Skip to content

docs: describe conditions and advice for 29 repository and framework skills - #285

Open
Dong1017 wants to merge 15 commits into
InferMatrixLab:mainfrom
Dong1017:codex/issue91-batch4
Open

docs: describe conditions and advice for 29 repository and framework skills#285
Dong1017 wants to merge 15 commits into
InferMatrixLab:mainfrom
Dong1017:codex/issue91-batch4

Conversation

@Dong1017

@Dong1017 Dong1017 commented Oct 31, 2026 •
edited
Loading

Copy link
Copy Markdown
Collaborator

Governing Issue and delivery slice

Refs #91, Round 2 Batch 4.

The requester authorized the parallel bounded continuation. Fill the four missing repository/framework applicability declarations, extend existing conditions where the bodies support them, and add procedure-specific advice for all 29 frozen IDs. Preserve the 25 existing applicability declarations. Unsupported refusal conditions are not invented.

Only applicability/recommendations change in the selected sidecars; source bodies, provenance, maturity and empirical-validation fields are unchanged.

The original frozen baseline is main ffb77f9a168ba1cb22f7ba261f289d24dbb6698f. Batch 2 is merged through #263 and retained in this integration; the other batches are separate deliveries. Shared projections must be regenerated from combined sources when integrating siblings, preserving all source and matrix additions. A closed applicability key allowlist remains deferred. No new source extraction, model/device execution, paid evaluation or efficacy claim is included.

The branch integrates main 919ed42c0d01dc0aa3b613085b9041d735d2d4db, retaining the Sarathi, DistServe and Mooncake sources, both matrices and upstream evidence changes.

Software consumers of the new metadata

  • Keep the broad MLX query-contribution and attribution checks, and add an independent semantic regression in both pytest and the installed CLI. A scoped query (mlx, lazy evaluation, repository inferrsi, scenario skills.discovery, limit 12) must retain the fixed MLX lifecycle abstract, its original document path and both query attributions. The expected identity is specified independently of search results, addressing the P2 coverage finding. Exact installed MLX resource-byte checks remain. Production search and its output limits are unchanged.
  • The full audit-vllm-kv-lifetimes K payload includes metadata, so this is a new payload identity. The 20261030 qualification plan, integrity manifest and historical receipts are preserved. A separate prospective 20261031 model-free plan pins the complete three-file owner at a04c599c49465b67ea40a1f0de622706ab3da955.
  • Ready selects that new plan for fresh exact-revision scripted software qualification. The production qualifier, complete-file/digest checks, task gates, 1,800-lease workload and limits remain unchanged. Live admission, paid authorization and reservations are unchanged; a software receipt cannot authorize a live attempt or establish knowledge efficacy.
  • The combined evidence inventory validates 21 bundles and 430 hashes with no errors, preserving main's new historical records and this branch's separate qualification plan. Unavailable external material remains explicitly unavailable; no historical manifest or live authorization is rewritten.

Checks and provenance

  • Current head: 680ae19b35e0dce8a5bc1d2d6941b995078e4668; integrated main: 919ed42c0d01dc0aa3b613085b9041d735d2d4db.
  • Reachable combined-source snapshot: 78002cb43fdea1956876880ab251f9f1cd9a367f. A separate following commit changes only the generated catalog and coverage; the final skill-library inputs are identical to this snapshot.
  • Three-way preservation checks retain the integrated main snapshot, the accepted batch fields, both matrices and historical evidence. Both native generator checks and full-tree Ruff pass.
  • The fixed MLX regression passed before integration; the combined evidence inventory and its targeted test pass. The broad Windows suite timed out without a completed verdict.
  • Linux Ready CI passes all six jobs on this head merged with main 919ed42c0d01dc0aa3b613085b9041d735d2d4db.

These are source and software checks. Independent review applies only to its recorded revision; this description is not an approval. Skill Eval execution, efficacy, native-device qualification and whole-tracker acceptance are not claimed. Refs #91; the tracker remains open.

Frozen IDs and field-to-source review pointers

Pointers refer to unchanged source sections in the canonical snapshot above. Existing applicability values are preserved.

Owner Field origins
audit-afd-p2p-exchanges applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:8-20,69-89 (Selected path; Rule 2)
applicability.input_contract: SKILL.md:35-42 (Required input)
applicability.produced_artifact: SKILL.md:90-103,113-118 (Rule 2; Verification proposal)
applicability.failure_modes: SKILL.md:58-65,76-93 (Rules 1 and 2)
applicability.refusal_point: SKILL.md:63-65 (Rule 1 stop)
recommendations: SKILL.md:44-61 (Rule 1); SKILL.md:76-84,97-103 (Rule 2; Verification proposal); SKILL.md:85-93 (Rule 2)
audit-flare-session-lifecycle applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:40-45,73-85 (Procedure; Actions, cost and stopping point)
applicability.input_contract: SKILL.md:30-38 (Inputs)
applicability.produced_artifact: SKILL.md:8-11,73-78 (Introduction; Actions, cost and stopping point)
applicability.failure_modes: SKILL.md:46-70 (Procedure)
applicability.refusal_point: SKILL.md:80-85 (Actions, cost and stopping point)
recommendations: SKILL.md:46-50 (Procedure step 2); SKILL.md:51-55,62-70 (Procedure steps 3, 5 and 6); SKILL.md:56-66 (Procedure steps 4 and 5)
audit-h3-metal-fasth3-semantics applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:8-15,60-76 (Introduction; Procedure step 6; Stop and report)
applicability.input_contract: SKILL.md:17-23 (Inputs and output)
applicability.produced_artifact: SKILL.md:25-27 (Inputs and output)
applicability.failure_modes: SKILL.md:31-59 (Procedure steps 1-5)
applicability.refusal_point: SKILL.md:69-76 (Stop and report)
recommendations: SKILL.md:31-36 (Procedure step 1); SKILL.md:37-46 (Procedure steps 2-3); SKILL.md:47-59 (Procedure steps 4-5)
audit-h3-runtime-residency applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:8-13,17-22,69-85 (Introduction; Inputs; Report and stop)
applicability.input_contract: SKILL.md:15-22 (Inputs and decision)
applicability.produced_artifact: SKILL.md:24-26,77-85 (Inputs and decision; Report and stop)
applicability.failure_modes: SKILL.md:24-26,38-48,62-68 (Decision; Procedure steps 2-3 and 6)
applicability.refusal_point: SKILL.md:77-85 (Report and stop)
recommendations: SKILL.md:30-37 (Procedure step 1); SKILL.md:43-54 (Procedure steps 3-4); SKILL.md:55-75 (Procedure steps 5-7)
audit-llama-runtime-state applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:3,16,55-57 (Description; source boundary; Stop conditions)
applicability.input_contract: SKILL.md:30-34 (Required inputs)
applicability.produced_artifact: SKILL.md:45-49 (Diagnostic procedure; Expected outputs)
applicability.failure_modes: SKILL.md:39-44 (Diagnostic procedure)
applicability.refusal_point: SKILL.md:55-57 (Stop conditions and trade-offs)
recommendations: SKILL.md:39-40 (Diagnostic procedure steps 2 and 3); SKILL.md:41-42 (Diagnostic procedure steps 4 and 5); SKILL.md:43-44 (Diagnostic procedure steps 6 and 7)
audit-recurrent-kv-finalization applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:10-16,47 (Select the supported path; Verify and recover)
applicability.input_contract: SKILL.md:12,23-30,45-47 (Supported path; Trace one token; Verify and recover)
applicability.produced_artifact: SKILL.md:23-30,47-49 (Trace one token; Verify and recover)
applicability.failure_modes: SKILL.md:18-21,25-41 (Two invariants; Trace; Decisive observations)
applicability.refusal_point: SKILL.md:43-49 (Verify and recover within scope)
recommendations: SKILL.md:20,27-29 (Two invariants; Trace steps 3-5); SKILL.md:20,27,38-39,45 (Invariants; Decisive observations; Verification); SKILL.md:21,29-30 (Invariants; Trace steps 5-6)
audit-sd-cpp-residency applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:8-23,57-61,72-83 (Introduction; Procedure steps 4 and 7-8)
applicability.input_contract: SKILL.md:25-36 (Required inputs)
applicability.produced_artifact: SKILL.md:79-92 (Bounded decision; Expected output)
applicability.failure_modes: SKILL.md:31-36,50-71 (Required inputs; Procedure steps 3-6)
applicability.refusal_point: SKILL.md:79-83 (Procedure step 8)
recommendations: SKILL.md:45-56 (Procedure steps 2-3); SKILL.md:57-71 (Procedure steps 4-6); SKILL.md:72-78 (Procedure step 7)
audit-sglang-diffusion-lifetimes applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:3,11-29,43-47 (Description; Boundary; Executor)
applicability.input_contract: SKILL.md:31-39 (Required inputs)
applicability.produced_artifact: SKILL.md:77-86 (Expected output and stop conditions)
applicability.failure_modes: SKILL.md:43-70 (Bounded procedure steps 1-5)
applicability.refusal_point: SKILL.md:77-86 (Expected output and stop conditions)
recommendations: SKILL.md:43-57 (Bounded procedure steps 1-3); SKILL.md:58-64,71-75 (Bounded procedure steps 4 and 6); SKILL.md:65-70 (Bounded procedure step 5)
audit-vllm-kv-lifetimes applicability.scope: SKILL.md:8-14,38-49 (Trigger and boundary; Required evidence; Decision procedure)
applicability.prerequisites: SKILL.md:3,12,38-40,61 (Description; Required evidence; Escalation)
applicability.input_contract: SKILL.md:38-40 (Required evidence)
applicability.produced_artifact: SKILL.md:49-53 (Decision procedure; Available actions)
applicability.failure_modes: SKILL.md:44-48,57 (Decision procedure; Counter-signals)
applicability.refusal_point: SKILL.md:55-61 (Counter-signals, verification, and escalation)
recommendations: SKILL.md:44-45,57 (Decision procedure steps 1-2; Counter-signals); SKILL.md:46-47,59 (Decision procedure steps 3-4; Verification); SKILL.md:48 (Decision procedure step 5)
bound-windowed-attention-replay applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: metadata.json#/applicability/prerequisites (preserved)
applicability.hardware: metadata.json#/applicability/hardware (preserved)
applicability.input_contract: SKILL.md:32-38 (Required inputs)
applicability.produced_artifact: SKILL.md:68-74 (Expected outputs)
applicability.failure_modes: SKILL.md:42-64,83-98 (Procedure; Caveats; Stop conditions)
applicability.refusal_point: SKILL.md:16-20,94-98 (Trigger and boundary; Stop conditions)
recommendations: SKILL.md:42-44,52-60 (Procedure steps 1-2 and 6-8); SKILL.md:58-62 (Procedure steps 7-9); SKILL.md:76-92 (Evidence to collect; Preconditions and caveats)
budget-speculative-verification applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: metadata.json#/applicability/prerequisites (preserved)
applicability.hardware: metadata.json#/applicability/hardware (preserved)
applicability.input_contract: SKILL.md:30-36 (Required inputs)
applicability.produced_artifact: SKILL.md:58-62 (Expected outputs)
applicability.failure_modes: SKILL.md:40-56,75-82 (Procedure; Preconditions and caveats)
applicability.refusal_point: SKILL.md:17-20,84-91 (Trigger and boundary; Stop conditions)
recommendations: SKILL.md:40-44 (Procedure steps 1-3); SKILL.md:46-56 (Procedure steps 4-8); SKILL.md:64-73 (Evidence to collect)
comfy-ui applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:8-11,22-31,53-63 (Introduction; Inputs; Procedure steps 5-6)
applicability.input_contract: SKILL.md:22-31 (Inputs)
applicability.produced_artifact: SKILL.md:70-76 (Output)
applicability.failure_modes: SKILL.md:35-58 (Procedure steps 1-5)
applicability.refusal_point: SKILL.md:64-68 (Procedure step 7)
recommendations: SKILL.md:38-47 (Procedure steps 2-3); SKILL.md:48-58 (Procedure steps 4-5); SKILL.md:59-63 (Procedure step 6)
dcp-async-save-performance applicability.scope: SKILL.md:3,14-20,36-61,96-107 (Description; Scope; Completion contract; Limits)
applicability.prerequisites: SKILL.md:10-12,22-34 (Status; Scope and decision order)
applicability.input_contract: SKILL.md:22-34,83-94 (Scope and decision order; Evidence to request)
applicability.produced_artifact: SKILL.md:36-81,83-94 (Completion contract; Performance lever; Evidence)
applicability.failure_modes: SKILL.md:26-32,47-61,100-102 (Scope; Completion contract; Limits)
applicability.refusal_point: SKILL.md:65-68,100-102 (Performance lever; Limits)
recommendations: SKILL.md:30-32 (Scope and decision order); SKILL.md:47-61,67-69 (Completion contract; Performance lever); SKILL.md:72-94 (Host budget; Evidence to request)
diagnose-vllm-admission-pressure applicability.scope: SKILL.md:8-14,44-51 (Trigger and boundary; Decision procedure)
applicability.prerequisites: SKILL.md:12,40-42 (Source boundary; Required evidence)
applicability.input_contract: SKILL.md:40-42 (Required evidence)
applicability.produced_artifact: SKILL.md:51-55 (Decision procedure; Available actions and costs)
applicability.failure_modes: SKILL.md:46-50,59 (Decision procedure; Counter-signals)
applicability.refusal_point: SKILL.md:57-63 (Counter-signals, verification, and escalation)
recommendations: SKILL.md:46-48 (Decision procedure steps 1-3); SKILL.md:49 (Decision procedure step 4); SKILL.md:50-51 (Decision procedure steps 5-6)
diffusers-invocation-state applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:8-10,31-33,41 (Boundary; Inputs; Procedure step 5)
applicability.input_contract: SKILL.md:31-33 (Inputs)
applicability.produced_artifact: SKILL.md:44-46 (Output and stop)
applicability.failure_modes: SKILL.md:37-41 (Procedure steps 1-5)
applicability.refusal_point: SKILL.md:44-46 (Output and stop)
recommendations: SKILL.md:37-39 (Procedure steps 1-3); SKILL.md:40-41 (Procedure steps 4-5); SKILL.md:42 (Procedure step 6)
dwarfstar applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:12-28,47-57,76-81 (Introduction; Routes; Native preparation; Stop)
applicability.input_contract: SKILL.md:35-45 (Establish the execution tuple)
applicability.produced_artifact: SKILL.md:76-81 (Stop and report)
applicability.failure_modes: SKILL.md:17-20,24-28,43-45,71-74 (Evidence boundary; Routes; Execution tuple; Hypothesis)
applicability.refusal_point: SKILL.md:76-81 (Stop and report)
recommendations: SKILL.md:22-28,63-67 (Route the question; Work one bounded hypothesis); SKILL.md:68-74 (Work one bounded hypothesis); SKILL.md:17-20,43-45,73-74 (Evidence boundary; Execution tuple; Hypothesis)
generation-cache-lifecycle applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:8-10,18-23 (Boundary; Procedure)
applicability.input_contract: SKILL.md:12-14 (Inputs)
applicability.produced_artifact: SKILL.md:25-27 (Output and stop conditions)
applicability.failure_modes: SKILL.md:18-22 (Procedure steps 1-5)
applicability.refusal_point: SKILL.md:25-27 (Output and stop conditions)
recommendations: SKILL.md:18 (Procedure step 1); SKILL.md:20 (Procedure step 3); SKILL.md:21-23 (Procedure steps 4-6)
husky applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:12-27,46-49,90-95 (Document boundary; Identity; Qualification limits)
applicability.input_contract: SKILL.md:29-40,46-49 (Establish the claim; Procedure step 2)
applicability.produced_artifact: SKILL.md:67-69,82-88 (Procedure step 8; Stop and report)
applicability.failure_modes: SKILL.md:37-40,46-59,77-80 (Evidence classes; Procedure; Offline support)
applicability.refusal_point: SKILL.md:67-69,82-88 (Procedure step 8; Stop and report)
recommendations: SKILL.md:29-40 (Establish the claim and evidence class); SKILL.md:44-55 (Procedure steps 1-4); SKILL.md:25-27,71-80 (Route the work; Safe offline support)
kv-tier-management applicability.scope: SKILL.md:3,10-30,34-44 (Description; Source scope; SearchSpace; Tier hierarchy)
applicability.prerequisites: SKILL.md:10-16,106-118,184-197 (Source scope; Lazy offload; Known limitations)
applicability.input_contract: SKILL.md:79-104,120-149,170-180,199-207 (Prefetch; Eviction; Composition; Measurement; Verification checklist)
applicability.produced_artifact: SKILL.md:46-83,161-180,199-207 (Policies; RAM-release levers; Measurement; Verification)
applicability.failure_modes: SKILL.md:79-104,184-207 (Prefetch; Eviction; Known limitations; Verification)
recommendations: SKILL.md:46-65,85-98 (Store policies; Retention; Eviction); SKILL.md:67-83 (Prefetch); SKILL.md:94-104,170-180,199-207 (Eviction; Measurement; Verification checklist)
lazy-evaluation-stream-lifetimes applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:8-14,23-29 (Boundary; Inputs; Validation; Output)
applicability.input_contract: SKILL.md:12-14 (Inputs)
applicability.produced_artifact: SKILL.md:25-27 (Output and stop conditions)
applicability.failure_modes: SKILL.md:18-22,29 (Procedure steps 1-5; Claim boundary)
applicability.refusal_point: SKILL.md:25-27 (Output and stop conditions)
recommendations: SKILL.md:18-19 (Procedure steps 1-2); SKILL.md:20-21 (Procedure steps 3-4); SKILL.md:14,22-23 (Inputs; Procedure steps 5-6)
lily applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:12-27,40-51,61-78 (Introduction; Routes; Native preparation; Qualification boundary)
applicability.input_contract: SKILL.md:29-38 (Establish the execution tuple)
applicability.produced_artifact: SKILL.md:17-27,66-78 (Routes; Stop and report)
applicability.failure_modes: SKILL.md:21-27,36-38,55-64,74-78 (Routes; Execution tuple; Hypothesis; Qualification boundary)
applicability.refusal_point: SKILL.md:66-78 (Stop and report)
recommendations: SKILL.md:55-58 (Work one bounded hypothesis); SKILL.md:59-60 (Work one bounded hypothesis); SKILL.md:36-38,61-78 (Execution tuple; Hypothesis; Stop and report)
local-dream applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:12-28,30-36,80-95 (Boundary; Routes; Identities; Stronger evidence)
applicability.input_contract: SKILL.md:30-57 (Establish identities before suggesting a change)
applicability.produced_artifact: SKILL.md:80-95 (Evidence required for a stronger conclusion)
applicability.failure_modes: SKILL.md:37-57,74-78 (Identities; Offline limits)
recommendations: SKILL.md:23-24,41-48 (Routes; Establish identities step 3); SKILL.md:25-26,32-40,74-78 (Routes; Identities; Offline limits); SKILL.md:49-57 (Establish identities steps 4-5)
mistral-rs applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:12-30,49-59 (Scope; Routes; Evidence; Preparation)
applicability.input_contract: SKILL.md:32-37 (Establish identity before proposing a change)
applicability.produced_artifact: SKILL.md:77-84 (Return a bounded result)
applicability.failure_modes: SKILL.md:18-30,38-48 (Routes; Establish identity steps 2-4)
recommendations: SKILL.md:18-30,38-44 (Routes; Establish identity steps 2-3); SKILL.md:45-52 (Establish identity steps 4-5); SKILL.md:54-64 (Preparation and use boundary)
mlc-llm applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:12-16,50-63 (Boundary; Native preparation and evidence limits)
applicability.input_contract: SKILL.md:27-44 (Establish the actual boundary)
applicability.produced_artifact: SKILL.md:18-25,65-70 (Routes; Return boundary)
applicability.failure_modes: SKILL.md:29-44 (Establish the actual boundary)
recommendations: SKILL.md:29-36 (Establish the actual boundary steps 1-2); SKILL.md:37-40 (Establish the actual boundary step 3); SKILL.md:41-44 (Establish the actual boundary step 4)
mlx-swift-lm applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:12-30,47-50,52-73 (Boundary; Routes; Preparation)
applicability.input_contract: SKILL.md:32-43 (Establish identity and evidence)
applicability.produced_artifact: SKILL.md:75-83 (Return a bounded result)
applicability.failure_modes: SKILL.md:34-43,54-60 (Establish identity; Preparation boundary)
recommendations: SKILL.md:37-40 (Establish identity step 2); SKILL.md:41-43 (Establish identity step 3); SKILL.md:54-68 (Preparation boundary)
project-maya applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:12-31,45-64 (Boundary; Routes; Evidence requirements)
applicability.input_contract: SKILL.md:33-47 (Establish identity before changing a setting)
applicability.produced_artifact: SKILL.md:56-64 (Return a bounded conclusion)
applicability.failure_modes: SKILL.md:27-31,35-50 (Coverage; Establish identity)
applicability.refusal_point: SKILL.md:56-64 (Return a bounded conclusion)
recommendations: SKILL.md:18-31 (Route by the failing contract); SKILL.md:42-44 (Establish identity step 3); SKILL.md:45-54 (Establish identity steps 4-6)
splash applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:12-23,42-66,68-79 (Boundary; Routes; Native preparation; Hypothesis; Stop)
applicability.input_contract: SKILL.md:30-40 (Establish the execution tuple)
applicability.produced_artifact: SKILL.md:68-79 (Stop and report)
applicability.failure_modes: SKILL.md:38-40,63-66 (Execution tuple; Work a bounded hypothesis)
applicability.refusal_point: SKILL.md:68-79 (Stop and report)
recommendations: SKILL.md:17-23,57-60 (Routes; Work a bounded hypothesis); SKILL.md:38-40,61-66 (Execution tuple; Work a bounded hypothesis); SKILL.md:42-53 (Native preparation boundary)
vllm-metal applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:12-34,42-55 (Boundary; Routes; Establish identity and authority)
applicability.input_contract: SKILL.md:36-41 (Establish identity and authority)
applicability.produced_artifact: SKILL.md:74-82 (Return a bounded conclusion)
applicability.failure_modes: SKILL.md:22-34,42-51,69-82 (Routes; Identity; Preparation; Conclusion)
recommendations: SKILL.md:22-25,42-48,69-72 (Routes; Identity; Preparation); SKILL.md:29-34 (Owner boundaries); SKILL.md:49-55,74-82 (Identity and authority; Return a bounded conclusion)
lithos-metal applicability.scope: metadata.json#/applicability/scope (preserved)
applicability.prerequisites: SKILL.md:8-19,33-43,59-72 (Boundary; Package contract; Work sequence)
applicability.input_contract: SKILL.md:33-38,59-64 (Package contract; Work sequence)
applicability.produced_artifact: SKILL.md:21-31,59-72 (Choose a route; Work sequence)
applicability.failure_modes: SKILL.md:29-31,74-84 (Cross-route boundary; Counter-signals and limits)
recommendations: SKILL.md:29-31 (Choose a route); SKILL.md:59-69 (Work sequence); SKILL.md:16-19,67-80 (Boundary; Work sequence; Counter-signals)

Knowledge authorship declaration

{
  "schema_version": 1,
  "head_sha": "680ae19b35e0dce8a5bc1d2d6941b995078e4668",
  "human_only": false,
  "material_ai_authors": [
    {
      "tool": "Codex",
      "model_id": "gpt-6-astra",
      "effort": "xhigh",
      "role": "author"
    }
  ]
}

@Dong1017
docs: describe repository and framework skill conditions
a04c599 
Preserve source provenance and evidence states while making the declared scope explicit.

Agent-assisted-by: Codex / gpt-6-astra [effort=xhigh] (author)
Signed-off-by: Dong1017 <55723162+Dong1017@users.noreply.github.com>
@Dong1017
docs: regenerate skill projections from published source history
57bc805 
Keep the canonical source snapshot reachable as a preceding commit.

Agent-assisted-by: Codex / gpt-6-astra [effort=xhigh] (author)
Signed-off-by: Dong1017 <55723162+Dong1017@users.noreply.github.com>
@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Oct 31, 2026 •
edited
Loading

Copy link
Copy Markdown

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review ⚠️ Failed 2026-10-31T04:01:45.937015Z
We couldn’t complete this review. Comment @codex review to try again.		 57bc805 PR opened
ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

@github-actions

github-actions Bot commented Oct 31, 2026 •
edited
Loading

Copy link
Copy Markdown

Skill evaluation

PLANNED: not executed; there are no task outcomes yet.
No approval, model efficacy or current-head test result is implied.

  • Next: run the pinned source/index --check below at head 680ae19b35e0dce8a5bc1d2d6941b995078e4668.
  • Task selection: no supported automatic comparison; choose a relevant source/content check and disposition uncovered claims.
  • Selected comparisons: 0 · Estimated task calls: 0. Uncovered/unsupported changes: 120; Canonical library paths: 29; Generated projection paths: 78; Other paths: 13. These are paths, not independent claims.
  • Return: command, input/source pin, exit status, outputs/failures and untested claims for independent verified/refuted/unverified disposition.
Author verification steps (nonpaid; not yet run)
  1. Bind the checkout. Run git rev-parse HEAD; it must equal 680ae19b35e0dce8a5bc1d2d6941b995078e4668. If it differs, obtain a current plan; do not reuse old results.
  2. Check source/index consistency. At that PR checkout, using the reviewed repository tool and its documented development dependencies, run:
python tools/build_skill_index.py --source-revision 78002cb43fdea1956876880ab251f9f1cd9a367f --check

The declared source pin matches this head's canonical library bytes. The command has not run: retain its exit code and full drift/success output. It checks the generated projection, not claim truth or skill efficacy.
No supported automatic comparison is selected. Review the canonical changes and explain which claims need evidence, are unaffected, or remain unverified; do not treat generated records as separate behavioral claims.
Uncovered scope by location: Canonical library paths: 29; Generated projection paths: 78; Other paths: 13. Path counts are not claim counts; no metadata-only conclusion is inferred.
3. Return evidence for independent review: head, check/task/source pin, command, environment, exit status, outputs/failures and untested claims. Label scripted controls. Request each claim's disposition: verified, refuted or unverified; this plan proves neither model provenance nor benefit.
Check recipes and evidence limits

Evidence and logs · Retention: 30 days.
Retained artifacts are verifiable only before Unix time 1796185381 while the linked artifact is available. Expired evidence is unverified.

Plan bindings, budget, frozen exam and advisory details

Advisory: obtain missing evidence

Retained evidence availability: planned / unexecuted. No approval or merge authorization is issued.
Applicable only while these revisions and inputs match the current PR.
Paid execution requires trusted label authorization and cumulative PR reservation. Unknown accounting blocks new paid calls; planning is not execution.
Pipeline evidence only; offline fixed outputs do not establish skill efficacy.
Unchanged-correct is valid neutral evidence, not incremental benefit.
Content/source review and authorship admission remain separate.
Trusted regrading checks retained answers, not model/hardware provenance or a hostile executor; this is not an efficacy certification.

  • Retained evidence is planned / unexecuted.
  • No supported comparison cases were selected.
  • Uncovered/unsupported scope needs a case or explicit reviewer disposition.
    Plan SHA-256: 10fd71b260c86d4102206846d0a9f9f62337c42995781c858b4801fabca20d8e
    Base: 919ed42c0d01dc0aa3b613085b9041d735d2d4db · Head: 680ae19b35e0dce8a5bc1d2d6941b995078e4668
    Trusted implementation: 919ed42c0d01dc0aa3b613085b9041d735d2d4db

Model/settings: {"effort":"not_applicable","name":"claude-opus-5-5"}
Budget: {"comparisons":0,"estimated_calls":0,"limits_per_comparison":{"campaign_seconds":600,"max_task_calls":2,"repeats":1,"timeout_seconds":180,"token_stop_new_trials":200000},"provider_config_sha256":"b28d80f8194fbe0f5a9f01515efb24b218979fd45f702776560da276f4b30d04","reservation":null}

Skill Comparison Cases Estimated calls

Automatic execution / cumulative PR budget:

{"reason":"Trusted configuration or pricing missing, invalid or expired","remaining_usd":"unknown","reserved_or_unresolved_usd":"unknown","spent_or_legacy_usd":"unknown","state":"blocked"}

Projection check selection (not an executed result):

{"declared_in":"src/inferrsi/data/repo_packs/inferrsi/knowledge/README.md","source_revision":"78002cb43fdea1956876880ab251f9f1cd9a367f","status":"ready"}

Provider/settings: {"client":{"api_version":"2023-06-01","endpoint":"https://yibuapi.com/v1/messages","kind":"anthropic-messages-full-context-v1","max_output_tokens":4096,"max_read_bytes":32768,"max_request_bytes":262144,"max_requests":1,"max_tool_calls":12,"provider":"yibuapi"},"limits":{"token_stop_new_trials":200000},"model":{"effort":"not_applicable","name":"claude-opus-5-5"},"pricing":{"currency":"USD","expires_at":1793271228,"group":"default","max_input_tokens":1000000,"model":"claude-opus-5-5","rate_evidence":"https://github.com/InferMatrixLab/InferRSI/issues/215#issuecomment-6467531341","rates":{"cache_read":"0.4","cache_write":"10","input":"8","output":"40"},"request_fee":"0","unit":"per_million_tokens"},"spend_limit":{"amount":"200","currency":"USD"}}

Uncovered / unsupported paths (120)
  • uncovered: ".github/workflows/ci-ready.yml"
  • uncovered: "docs/skill-coverage.md"
  • uncovered: "evidence/kv-actor-qualification-20261031/README.md"
  • uncovered: "evidence/kv-actor-qualification-20261031/patch/README.md"
  • uncovered: "evidence/kv-actor-qualification-20261031/plan/integrity.json"
  • uncovered: "evidence/kv-actor-qualification-20261031/plan/qualification-plan.json"
  • uncovered: "evidence/kv-actor-qualification-20261031/results/README.md"
  • uncovered: "evidence/kv-actor-qualification-20261031/validator/README.md"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/README.md"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/ascend/skills/ascend-aclgraph-capture-replay.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/ascend/skills/ascend-device-health.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/ascend/skills/ascend-distributed-parallel.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/ascend/skills/ascend-graph-mode.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/ascend/skills/ascend-model-adaptation.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/ascend/skills/ascend-operator-profiling.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/ascend/skills/ascend-operator-tuning.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/ascend/skills/ascend-optimization-casebook.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/ascend/skills/ascend-performance-analysis.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/ascend/skills/ascend-video-diffusion-optimization.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/ascend/skills/ascendc-tiling-design.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/ascend/skills/cann-attribution-overlap.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/ascend/skills/torchnpu-device-memory-lifecycle.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/ascend/skills/torchnpu-pinned-host-memory-lifecycle.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/assess-serving-preemption.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/assess-split-backward-schedule.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/audit-fixed-workload-claim.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/audit-h3-metal-fasth3-semantics.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/audit-h3-runtime-residency.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/audit-llama-runtime-state.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/audit-recurrent-kv-finalization.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/audit-sd-cpp-residency.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/audit-sglang-diffusion-lifetimes.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/audit-vllm-kv-lifetimes.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/bound-windowed-attention-replay.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/budget-speculative-verification.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/calibrate-serving-simulation.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/cgroup-resource-controls.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/checkpoint-recovery-consistency.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/choose-parallel-training-plan.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/cite-check-downstream-pr.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/claim-contract.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/comfy-ui.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/competitor-scan.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/dcp-async-save-performance.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/dense-performance-evidence.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/diagnose-vllm-admission-pressure.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/diffusers-invocation-state.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/draft-acceptance-issue.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/dwarfstar.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/evaluate-elastic-serving-placement.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/generation-cache-lifecycle.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/host-memory-management.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/host-overhead-optimization.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/husky.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/inference-benchmark-design.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/inference-optimization-casebook.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/kernel-writing-guide.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/kv-tier-management.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/lazy-evaluation-stream-lifetimes.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/lily.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/lithos-metal.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/local-dream.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/mistral-rs.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/mlc-llm.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/mlx-swift-lm.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/npu-skill-factory.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/optimization-search-policy.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/os-process-placement.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/performance-ceiling-analysis.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/plan-activation-residency.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/project-maya.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/repository-evolution-analysis.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/rsi-evidence.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/skill-evolution-playbook.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/splash.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/storage-io-measurement.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/training-input-pipeline.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/training-resource-budget.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/video-diffusion-optimization.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/vllm-metal.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/common/skills/weight-loading-optimization.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/cuda/skills/audit-afd-p2p-exchanges.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/cuda/skills/audit-flare-session-lifecycle.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/cuda/skills/diagnose-attention-bottlenecks.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/cuda/skills/host-overhead-profiling.json"
  • uncovered: "src/inferrsi/data/repo_packs/inferrsi/knowledge/cuda/skills/reason-hopper-kernels.json"
  • uncovered: "src/inferrsi/data/skill_library/frameworks/diffusers/diffusers-invocation-state/metadata.json"
  • uncovered: "src/inferrsi/data/skill_library/frameworks/mlx/lazy-evaluation-stream-lifetimes/metadata.json"
  • uncovered: "src/inferrsi/data/skill_library/frameworks/pytorch/dcp-async-save-performance/metadata.json"
  • uncovered: "src/inferrsi/data/skill_library/repositories/afd-plugin/audit-afd-p2p-exchanges/metadata.json"
  • uncovered: "src/inferrsi/data/skill_library/repositories/comfy-ui/metadata.json"
  • uncovered: "src/inferrsi/data/skill_library/repositories/deepinfer/bound-windowed-attention-replay/metadata.json"
  • uncovered: "src/inferrsi/data/skill_library/repositories/deepinfer/budget-speculative-verification/metadata.json"
  • uncovered: "src/inferrsi/data/skill_library/repositories/dwarfstar/metadata.json"
  • uncovered: "src/inferrsi/data/skill_library/repositories/flare-local-inference/audit-flare-session-lifecycle/metadata.json"
  • uncovered: "src/inferrsi/data/skill_library/repositories/h3-c/audit-h3-runtime-residency/metadata.json"
  • uncovered: "src/inferrsi/data/skill_library/repositories/h3-metal/audit-h3-metal-fasth3-semantics/metadata.json"
  • uncovered: "src/inferrsi/data/skill_library/repositories/husky/metadata.json"
  • uncovered: "src/inferrsi/data/skill_library/repositories/lily/metadata.json"
  • uncovered: "src/inferrsi/data/skill_library/repositories/lithos-metal/metadata.json"
    Additional paths omitted; see retained evidence.

@Dong1017 Dong1017 added the area:knowledge Knowledge line: packs, skill library, knowledge admission label Oct 31, 2026 — with ChatGPT Codex Connector
Dong1017
Dong1017 commented Oct 31, 2026
View reviewed changes

@Dong1017 Dong1017 left a comment

Copy link
Copy Markdown
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not ready at 57bc805d: current Ready checks expose two concrete integration regressions.

  • P2 — The metadata change invalidates the frozen K input. repositories/vllm/audit-vllm-kv-lifetimes/metadata.json is part of the complete actor knowledge payload. Actor qualification now stops in tools/qualify_kv_edit_actor.py:52 with Pinned K bytes changed, before it can produce an actor receipt. Coordinate an explicit model-free qualification input/version for the changed knowledge, or read the original frozen K for that original plan. Preserve the historical plan, first attempt and authorization; changing its digest to force a pass or replaying the live pair would not resolve this correctly.
  • P2 — Installed research discovery regresses. The existing knowledge research MiniMax-H3 --framework mlx … --limit 12 check no longer returns inferrsi.common.skills.lazy_evaluation_stream_lifetimes and fails at tests/check_installed.py:296 on both Python versions. Descriptive sidecars feed the searchable generated JSON, so this change affects discovery ordering. Restore or explicitly reconcile that consumer contract and cover the enlarged metadata in its regression check; do not just remove the assertion to green CI.
  • Next action stays with the existing PR author, @Dong1017. Recheck the corrected head and its complete source projections before requesting LGTM. This review covers the two observed failures, not a completed semantic approval of all 29 sidecars.
Evidence and limits

Head: 57bc805db9bb56995e2ae988830e1c52a14f85d3; API base / assessed main: ecca7e9136d962747f42860f58e332f45b0aa358 / ecca7e9136d962747f42860f58e332f45b0aa358.
Ready run 41166742778: installed Python 3.10 failure, Python 3.11 failure, actor qualification failure. The missing actor upload is downstream of the K check, not proof that a model ran. checked_plan compares current owner bytes with the frozen three-file manifest; the plan pins the old metadata SHA-256 b43fc3a7ccae59c664fcdf9ff77962edcc07c2a3df8b59324db551642bd23e23.
The installed query's fixed result is asserted in check_installed.py. catalog search scores the serialized record, and research selection consumes those ordered results.
I inspected source and hosted logs; I did not rerun tests, workflows, models or devices. Declaration compliance alone does not clear these failures. COMMENTED because the posting account is also the PR author; no formal self-REQUEST_CHANGES or approval is issued.

This was referenced Oct 31, 2026
Open
Open
@Dong1017
ci: qualify the new knowledge payload without rewriting historical ev…
7b8995d 
…idence

Keep model-free qualification on a separately frozen current payload. Test framework-query contribution independently of incidental corpus ranking.

Agent-assisted-by: Codex / gpt-6-astra [effort=xhigh] (author)
Signed-off-by: Dong1017 <55723162+Dong1017@users.noreply.github.com>
@Dong1017
Merge main and retain the batch 4 knowledge source snapshot
e30ccae 
Preserve upstream Sarathi source, matrix and evidence additions alongside this batch. Regenerate shared projections in the following commit.

Agent-assisted-by: Codex / gpt-6-astra [effort=xhigh] (author)
Signed-off-by: Dong1017 <55723162+Dong1017@users.noreply.github.com>
@Dong1017
docs: regenerate skill projections from published source history
40db800 
Keep the canonical source snapshot reachable as a preceding commit.

Agent-assisted-by: Codex / gpt-6-astra [effort=xhigh] (author)
Signed-off-by: Dong1017 <55723162+Dong1017@users.noreply.github.com>
@Dong1017
Merge main and retain the batch 4 knowledge source snapshot
224f653 
Preserve upstream source, matrix and evidence additions alongside this batch. Regenerate shared projections in the following commit.

Agent-assisted-by: Codex / gpt-6-astra [effort=xhigh] (author)
Signed-off-by: Dong1017 <55723162+Dong1017@users.noreply.github.com>
@Dong1017
docs: regenerate skill projections from published source history
ca893f2 
Keep the canonical source snapshot reachable as a preceding commit.

Agent-assisted-by: Codex / gpt-6-astra [effort=xhigh] (author)
Signed-off-by: Dong1017 <55723162+Dong1017@users.noreply.github.com>
@Dong1017
Merge main Mooncake knowledge into Batch 4 repository guidance
a7f23df 
Preserve all 29 applicability and recommendation declarations and the
prospective model-free K qualification bundle and software consumers.
Retain upstream canonical sources and matrices before regeneration.

Agent-assisted-by: Codex / gpt-6-astra [effort=xhigh] (author)
Signed-off-by: Dong1017 <55723162+Dong1017@users.noreply.github.com>
@Dong1017
docs: regenerate skill projections from published source history
6b8093d 
Keep the canonical source snapshot reachable as a preceding commit.

Agent-assisted-by: Codex / gpt-6-astra [effort=xhigh] (author)
Signed-off-by: Dong1017 <55723162+Dong1017@users.noreply.github.com>
Dong1017
Dong1017 commented Oct 31, 2026
View reviewed changes

@Dong1017 Dong1017 left a comment

Copy link
Copy Markdown
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Incremental review of the current-head repair for the previously reported installed-research regression. Current Ready CI is green and the new pinned-K qualification plan is distinct from the preserved historical plan; I am not reopening those former failures. One new test-contract gap remains in the replacement MLX assertion (inline): passing now proves query propagation, not that a relevant MLX lifecycle resource survives discovery. Scope: static inspection of research.py, catalog.py, the changed research/installed tests and the existing review; no tests, models or hardware were run. This COMMENT is not merge or efficacy approval.

Comment thread tests/test_research.py
assert len(identities) == len(set(identities)) <= 12
assert "inferrsi.common.skills.lazy_evaluation_stream_lifetimes" in identities
# Corpus additions can reorder matches; the independent framework query still contributes.
framework_first = search_knowledge("mlx", limit=1)[0]["id"]

@Dong1017 Dong1017 Oct 31, 2026

Copy link
Copy Markdown
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

[P2, regression-test coverage] The replacement expected ID is computed with search_knowledge("mlx", limit=1) from the same mutable corpus under test. research_knowledge round-robins that stream and attaches its query attribution, so this check can pass even if new metadata overweights an unrelated record that merely repeats “mlx”; it no longer detects the disappearance of lazy_evaluation_stream_lifetimes that prompted the previous P2. This is a coverage risk, not evidence that the current runtime returns a wrong record. Please keep the dynamic round-robin/wiring assertion, but add a separate semantic regression with an independently specified expected MLX framework/resource identity or source-bound properties in a deliberately scoped query/limit, so corpus growth cannot make the oracle track the regression. The installed-package counterpart at tests/check_installed.py:296 has the same self-referential expectation.

@Dong1017 Dong1017 Nov 2, 2026

Copy link
Copy Markdown
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Fixed for this regression-test finding at 680ae19b.

  • The new scoped pytest and installed-CLI cases specify the MLX lifecycle ID and original document path independently of the search results, and require both framework and mechanism attribution. They cover the semantic identity that the broad wiring assertion alone could not protect.
  • Current-head Ready checks pass. I inspected the changed assertions and their query inputs; I did not rerun the suite locally.
  • This clears this test-contract finding only. It is not LGTM for all 29 sidecars; that source-fidelity review remains pending with this review task. The existing PR owner is unchanged, and no thread is marked resolved here.
Version and evidence

Reviewed head: 680ae19b35e0dce8a5bc1d2d6941b995078e4668; API base and assessed main: 919ed42c0d01dc0aa3b613085b9041d735d2d4db. Changed cases: tests/test_research.py::test_scoped_mlx_research_retains_the_lifecycle_resource and the scoped CLI block in tests/check_installed.py. Ready CI is software evidence; Skill Eval remains planned with zero comparisons and no efficacy result. No model/device or paid execution.

@Dong1017
test: retain independent MLX lifecycle discovery coverage
ed5d907 
Keep query propagation checks and separately require the canonical MLX lifecycle resource in a scoped research query, including installed CLI discovery.

Agent-assisted-by: Codex / gpt-6-astra [effort=xhigh] (author)
Signed-off-by: Dong1017 <55723162+Dong1017@users.noreply.github.com>
@Dong1017
Merge current main while preserving repository skill sources
daa33a4 
Retain both qualification inventories and all accepted method metadata; regenerate discovery views from this combined source snapshot separately.

Agent-assisted-by: Codex / gpt-6-astra [effort=xhigh] (author)
Signed-off-by: Dong1017 <55723162+Dong1017@users.noreply.github.com>
@Dong1017 Dong1017 mentioned this pull request Nov 1, 2026
Open
15 tasks
@Dong1017
docs: regenerate combined skill discovery projections
9e9f6b7 
Keep the canonical combined-source snapshot reachable in the preceding commit.

Agent-assisted-by: Codex / gpt-6-astra [effort=xhigh] (author)
Signed-off-by: Dong1017 <55723162+Dong1017@users.noreply.github.com>
@Dong1017 Dong1017 mentioned this pull request Nov 2, 2026
Open
9 tasks
@Dong1017
Merge current main while preserving Batch 4 source metadata
78002cb 
Resolve generated discovery conflicts provisionally; regenerate from this combined source snapshot in the following commit.

Agent-assisted-by: Codex / gpt-6-astra [effort=xhigh] (author)
Signed-off-by: Dong1017 <55723162+Dong1017@users.noreply.github.com>
@Dong1017
docs: regenerate Batch 4 discovery projections after main merge
a452bf8 
Bind discovery records to combined source snapshot 78002cb.

Agent-assisted-by: Codex / gpt-6-astra [effort=xhigh] (author)
Signed-off-by: Dong1017 <55723162+Dong1017@users.noreply.github.com>
@Dong1017
Merge H3 evidence while retaining actor campaign inventory
680ae19 
Preserve both evidence bundles and validate the combined 21-bundle inventory with 430 checked hashes. Skill library and generated discovery projections remain unchanged.

Agent-assisted-by: Codex / gpt-6-astra [effort=xhigh] (author)
Signed-off-by: Dong1017 <55723162+Dong1017@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

area:knowledge Knowledge line: packs, skill library, knowledge admission

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@Dong1017