Repository navigation
pm context --depth full + --parent subtree scope; pm list --no-truncate (GH-156, GH-154, GH-155, GH-223) - #234
Conversation
…ate (GH-156, GH-154) Context-retrieval ergonomics serving the core "project management = context management" philosophy: surface richer project context in fewer agent calls. pm context (GH-156 / pm-j0vc, pm-ds0m): - --depth full: comprehensive snapshot — every known section (ignoring the per-section settings toggles brief/standard/deep respect) with no per-section row cap (overridable via an explicit --limit). New depth in CONTEXT_DEPTH_VALUES. - --parent <id>: scope the focus items, hierarchy, agenda, and all derived sections to the anchor item plus its transitive descendants — the "status of this epic?" view for large trackers. Reuses the existing buildChildrenByParent/collectDescendants helpers (no duplicate tree walk); case-insensitive anchor match; unknown anchor → NOT_FOUND. filters.parent is echoed and the markdown header gains a "scope: subtree of <id>" line. pm list (GH-154 / pm-0c0j): - --no-truncate (alias --all): force the full matched set, overriding any --limit. - result.total: pre-pagination match count, emitted only when --limit/--offset actually dropped rows, in both compact and verbose paths. Already-shipped, verified + documented: - GH-155 / pm-gcm3: pm get child rollup for Milestone/Epic (count/active/by_status). - GH-223: --include-body pulls item bodies in bulk in one call (compact JSON stays token-light by default); documented prominently in COMMANDS.md / AGENT_GUIDE.md. Wiring: types, register-list-query options, normalize{List,Context}Options, CONTEXT/LIST flag contracts + commander-string-option contracts, MCP pm_context description, zsh/fish completions, contracts golden fixture, docs. Unit + integration tests for every new branch; full coverage gate green. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
![sourcery-ai[bot]](https://avatars.githubusercontent.com/in/48477?s=60&v=4){kind=small}
|
@codex review |
|
Warning Review limit reached
More reviews will be available in 40 minutes and 9 seconds. Learn how PR review limits work. Your organization has used up its prepaid credits, and credit purchases are no longer available. Enable the review add-on in the billing tab to keep reviews running — you're only billed for reviews past your plan's rate limits ($0.25/file). ⌛ How to resolve this issue?After more reviews become available, a review can be triggered using the We recommend that you space out your commits to avoid hitting the rate limit. 🚦 How do rate limits work?CodeRabbit enforces hourly rate limits for each developer per organization. Our paid plans include higher PR review limits than trial, open-source, and free plans. In all cases, reviews become available again over time. During sustained high-volume PR review activity, CodeRabbit may temporarily slow when the next review becomes available. Please see our Fair Usage Limits Policy for further information. ℹ️ Review info⚙️ Run configurationConfiguration used: Repository UI (base), Organization UI (inherited) Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (2)
📝 WalkthroughWalkthroughAdds three new Changespm list --no-truncate/--all, pm context --depth full, and --parent subtree scoping
Sequence Diagram(s)sequenceDiagram
participant User
participant CLI as CLI (register-list-query)
participant normalizeListOptions
participant runList as runList (list.ts)
participant Result as ListResult
User->>CLI: pm list --no-truncate --limit 5
CLI->>normalizeListOptions: truncate=false, limit=5
normalizeListOptions->>runList: noTruncate=true, limit=5
runList->>runList: noTruncate overrides limit → fetch all
runList->>runList: compute totalMatched
runList-->>Result: count=all_matched, no total field
Result-->>User: full result set
sequenceDiagram
participant User
participant CLI as CLI (register-list-query)
participant normalizeContextOptions
participant runContext as runContext (context.ts)
participant collectSubtreeIds
participant Corpus as Item Corpus + Agenda
User->>CLI: pm context --parent epic-1 --depth deep
CLI->>normalizeContextOptions: parent="epic-1", depth="deep"
normalizeContextOptions->>runContext: parentScope="epic-1"
runContext->>Corpus: fetch full item corpus
runContext->>collectSubtreeIds: "epic-1" + fullItems
collectSubtreeIds-->>runContext: subtreeIds (transitive)
alt anchor not found
runContext-->>User: PmCliError NOT_FOUND
end
runContext->>Corpus: filter items to subtreeIds
runContext->>Corpus: filter agenda.events to subtreeIds
runContext-->>User: snapshot with "scope: subtree of epic-1"
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Possibly related issues
Possibly related PRs
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
|
To use Codex here, create a Codex account and connect to github. |
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request introduces several context-retrieval and list ergonomics improvements to the pm CLI. Key additions include the pm context --depth full option to provide a comprehensive snapshot of all sections without per-section row caps, and pm context --parent <id> to scope the context snapshot to a single epic or parent subtree. Additionally, it adds pm list --no-truncate (alias --all) to override configured limits and surface a pre-pagination total count when results are truncated. The changes are accompanied by updated documentation, shell completions, MCP tool definitions, CLI contracts, and comprehensive unit tests. There are no review comments, so no further feedback is provided.
Important
The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.
There was a problem hiding this comment.
Code Review
This pull request introduces several context-retrieval ergonomics improvements to the pm CLI, including a new --depth full option for pm context to retrieve a comprehensive snapshot without per-section caps, a --parent <id> option to scope snapshots to a single epic/parent subtree, and a --no-truncate (alias --all) flag for pm list to override limits and return all matched rows. Additionally, it surfaces pre-pagination totals when results are truncated. The feedback suggests defensively checking that item.id is a string before calling .trim() in collectSubtreeIds to prevent potential runtime crashes on malformed user-defined front matter.
Important
The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.
| // section relies on, so subtree membership stays consistent across the snapshot. | ||
| function collectSubtreeIds(corpus: ItemFrontMatter[], parentId: string): { ids: Set<string>; found: boolean } { | ||
| const target = parentId.trim().toLowerCase(); | ||
| const anchor = corpus.find((item) => item.id.trim().toLowerCase() === target); |
There was a problem hiding this comment.
Defensively check that item.id is a string before calling .trim() on it. Since the CLI parses user-defined front matter from markdown files, a malformed file could have a missing or non-string id field, which would cause a runtime TypeError and crash the command.
| const anchor = corpus.find((item) => item.id.trim().toLowerCase() === target); | |
| const anchor = corpus.find((item) => typeof item.id === "string" && item.id.trim().toLowerCase() === target); |
There was a problem hiding this comment.
Declining: ItemFrontMatter.id is a non-optional string (src/types.ts:128) — the primary key derived from the item filename by the store loader, not arbitrary user front matter. It is consumed unguarded throughout this file (toContextFocusItem line 541, collectDescendants line 554, itemMap.set line 608, buildWorkload line 794/813), so a guard only in collectSubtreeIds would be inconsistent and wouldn't prevent a downstream crash anyway. Adding it would introduce an unreachable branch. The user-supplied --parent value is already trimmed/validated in parseContextParent.
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 5
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In @.agents/pm/features/pm-j0vc.toon:
- Line 3: Update the description field in this closed feature record to remove
the reference to the split-off `--fields` scope. The feature is closed as `pm
context --depth full` only, while the `--fields` functionality has been moved to
pm-hnjf. Edit the description to only describe what was shipped in this scope
(depth=full for context output and GitHub issue `#156`), removing any mention of
`--fields` customization. Apply the same cleanup to line 12 where this
description also appears.
In `@docs/COMMANDS.md`:
- Around line 120-125: The documentation in both files incorrectly states that
pm list* returns every matched row when no --limit is set, but this omits that
--offset alone also causes truncation. In docs/COMMANDS.md lines 120-125, change
the opening sentence to state "when neither --limit nor --offset is set" instead
of "when no --limit is set". Apply the identical wording change to
docs/AGENT_GUIDE.md lines 120-125 in the token-minimal retrieval note to ensure
both files accurately describe pagination behavior.
In `@src/cli/commands/context.ts`:
- Around line 1175-1178: When parent subtree scoping is needed (when
needsAllItems or parentScope is defined), the code must retrieve the complete
set of items without pagination constraints to accurately resolve parent
relationships. Currently, the allListOptions object inherits pagination controls
from baseListOptions, causing runList to return truncated results which leads to
incomplete or incorrect parent scope resolution. Remove pagination controls
(limit and offset) from allListOptions when parent scoping is required so that
the full item set is available for parent subtree evaluation. This same issue
applies at the other affected locations (around lines 1183-1190 and 1220-1231)
where agenda.events are similarly built with inherited pagination controls that
need to be disabled when parent scoping is active.
In `@tests/fixtures/contracts/full.json`:
- Around line 1601-1603: The schema definitions for action "context" and action
"ctx" in the contract fixture file need to be updated to match the newly added
command flags. For both the action: "context" and action: "ctx" schema blocks
(at lines 1601-1603 and also at lines 1990-1992 and 5898-5903), add the parent
property to the schema and update the depth enum constraint to include "full" in
addition to the existing "brief", "standard", and "deep" values. With
additionalProperties set to false, these schema changes are necessary to make
the new --parent flag and depth=full option valid according to the contract.
- Around line 2795-2800: The list-family action schemas in full.json have strict
validation (additionalProperties: false) but are missing the `--no-truncate` and
`--all` flag definitions that are present in the command_flags section. Add the
`--no-truncate` flag with alias `--all` to the schema definitions for all eight
list-family action blocks (list, list-all, list-draft, list-open,
list-in-progress, list-blocked, list-closed, list-canceled) by including the
flag object with its aliases in each action's schema oneOf block, matching the
structure shown in the command_flags section.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Repository UI (base), Organization UI (inherited)
Review profile: CHILL
Plan: Pro
Run ID: f3828979-1ec2-43de-a631-b1168571004c
📒 Files selected for processing (29)
.agents/pm/extensions/.managed-extensions.json.agents/pm/features/pm-0c0j.toon.agents/pm/features/pm-ds0m.toon.agents/pm/features/pm-gcm3.toon.agents/pm/features/pm-hnjf.toon.agents/pm/features/pm-j0vc.toon.agents/pm/history/pm-0c0j.jsonl.agents/pm/history/pm-ds0m.jsonl.agents/pm/history/pm-gcm3.jsonl.agents/pm/history/pm-hnjf.jsonl.agents/pm/history/pm-j0vc.jsonl.agents/pm/history/pm-kp1d.jsonl.agents/pm/stories/pm-kp1d.toonCHANGELOG.mddocs/AGENT_GUIDE.mddocs/COMMANDS.mdsrc/cli/commands/completion.tssrc/cli/commands/context.tssrc/cli/commands/list.tssrc/cli/register-list-query.tssrc/cli/registration-helpers.tssrc/mcp/tool-definitions.tssrc/sdk/cli-contracts.tssrc/sdk/cli-contracts/commander-types.tssrc/types.tstests/fixtures/contracts/full.jsontests/unit/context-command.spec.tstests/unit/list-command.spec.tstests/unit/registration-helpers.spec.ts
| id: pm-j0vc | ||
| title: pm context --depth full and --fields projection customization (GH-156) | ||
| title: "pm context --depth full: comprehensive snapshot (all sections, no per-section cap) (GH-156)" | ||
| description: "Add depth=full (all items with status) and --fields to customize per-row projection in context output. GitHub issue #156. Related: pm-ds0m (subtree scoping)." |
There was a problem hiding this comment.
Drop the split-off --fields scope from this closed feature.
This record is closed as pm context --depth full, but the description still claims --fields belongs to GH-156. Since pm-hnjf now owns that follow-up, the closed feature metadata should only describe the shipped scope.
Proposed cleanup
-description: "Add depth=full (all items with status) and --fields to customize per-row projection in context output. GitHub issue `#156`. Related: pm-ds0m (subtree scoping)."
+description: "Add depth=full (all items with status) in context output. GitHub issue `#156`. Related: pm-ds0m (subtree scoping)."Also applies to: 12-12
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In @.agents/pm/features/pm-j0vc.toon at line 3, Update the description field in
this closed feature record to remove the reference to the split-off `--fields`
scope. The feature is closed as `pm context --depth full` only, while the
`--fields` functionality has been moved to pm-hnjf. Edit the description to only
describe what was shipped in this scope (depth=full for context output and
GitHub issue `#156`), removing any mention of `--fields` customization. Apply the
same cleanup to line 12 where this description also appears.
… full/no-truncate in strict action schemas (CodeRabbit) Addresses CodeRabbit review on PR #234: - MAJOR: --parent subtree scoping ran against paginated reads. The structural corpus fetches (allItems, the active set when scoping, and the agenda) still inherited --limit/--offset, so `pm context --parent X --limit N` could resolve the subtree against a truncated corpus → false NOT_FOUND for an anchor that paginated out of view, plus incomplete agenda. --limit/--offset are per-section display caps (applied later via .slice), not corpus filters, so the structural reads now run unpaginated; the agenda pulls the full set before subtree filter. - MAJOR: strict MCP action schemas (additionalProperties:false) rejected the new options. Added `parent` to TOOL_CONTEXT_OPTION_CONTRACTS, `noTruncate` to the list-family contract keys + parameter property/metadata, and `full` to the shared `depth` enum (+ description). Regenerated the contract golden fixture. - Docs: clarify list returns every row only when neither --limit nor --offset is set (offset alone also truncates). - pm-j0vc description: drop the split-off --fields scope (now owned by pm-hnjf). Coverage gate green (97.7/91.7/97.7/98); static/contracts/docs-links gates pass. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
|
Thanks @coderabbitai — all findings addressed in e415062:
For the Gemini @coderabbitai review |
|
The Kicking off a fresh review of the updated branch now. ✅ Action performedReview finished.
|
|
To use Codex here, create a Codex account and connect to github. |
There was a problem hiding this comment.
Code Review
This pull request introduces several context-retrieval ergonomics improvements, including a new --depth full option for pm context to provide a comprehensive snapshot without per-section caps, a --parent <id> option to scope context snapshots to a single subtree, and a --no-truncate (alias --all) flag for pm list to override limits and return all matched rows. It also updates shell completions, MCP tool definitions, and documentation. A critical issue was identified in the implementation of --parent where filtering allItems globally to the subtree will break metadata resolution for blockers/dependencies outside the subtree and cause incorrect unparented/hygiene warnings; it is recommended to keep allItems unfiltered as a reference corpus and only filter a copy for specific rollups.
Important
The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.
| listedFrontMatter = listedFrontMatter.filter((item) => subtreeIds?.has(item.id)); | ||
| allItems = allItems.filter((item) => subtreeIds?.has(item.id)); |
There was a problem hiding this comment.
Filtering allItems globally to only contain subtree items will cause issues in other sections of the context snapshot:
- Blockers/Dependencies: Any blocker or dependency pointing to an item outside the subtree will fail to resolve its metadata (title, status) because that item is no longer in
allItems. This results in them being rendered asunknown <id>. - Unparented/Hygiene Checks: The anchor item (or other top-level items in the subtree) may be incorrectly flagged as 'unparented' or having invalid parent links because their parents (which lie outside the subtree) are filtered out of
allItems.
Suggestion: Keep the full, unfiltered allItems as the reference corpus for resolving item metadata (like blockers and parent links), and only filter a copy of it (e.g., subtreeItems) when passing it to sections that specifically build hierarchical or progress rollups (like hierarchy or progress).
There was a problem hiding this comment.
Fixed in f9f9a54. Point 1 (blocker cross-references) was valid: itemMap now builds from the unfiltered fullCorpus so a subtree item blocked by an item outside the subtree keeps its blocked_by title/status; only the enumeration sets stay subtree-scoped. Added an integration test verifying the cross-subtree blocker resolves. Point 2 (anchor flagged unparented) does not occur — buildUnparented only flags items with no parent (!normalizedParentId(item.parent)), and the anchor has a parent.
…us (Gemini) Addresses Gemini high-priority review on PR #234: filtering allItems globally to the subtree dropped cross-subtree blocker references. A subtree item blocked by an item OUTSIDE the subtree lost its blocked_by title/status (rendered as a bare id), because itemMap was built from the subtree-filtered set. Keep the unfiltered corpus (`fullCorpus`) as the reference for cross-item metadata resolution and build itemMap from it, while the enumeration sets (focus/hierarchy/ progress/recently_created/unparented/workload/staleness) stay subtree-scoped. (Gemini's second point — anchor flagged "unparented" — does not occur: buildUnparented only flags items with no parent, and the anchor has one.) Adds integration coverage for cross-subtree blocker resolution and for the unpaginated-corpus fix (tight --limit no longer triggers a false NOT_FOUND). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
|
Pushed f9f9a54 addressing the Gemini high-priority finding (cross-subtree blocker metadata now resolves against the full corpus). Re-requesting review. @coderabbitai review |
|
✅ Action performedReview finished.
|
There was a problem hiding this comment.
Code Review
This pull request introduces several context-retrieval ergonomics improvements, including the --depth full option for pm context to retrieve a comprehensive snapshot without per-section caps, the --parent <id> option to scope context snapshots to a single subtree, and the --no-truncate (alias --all) flag for pm list to override limits and surface pre-pagination totals. It also updates documentation, shell completions, MCP tool definitions, and unit tests. The review feedback highlights a critical improvement opportunity to ensure robust case-insensitive matching by normalizing and trimming IDs to lowercase when building and querying the subtreeIds set across collectSubtreeIds, list filters, and agenda event filters.
Important
The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.
| function collectSubtreeIds(corpus: ItemFrontMatter[], parentId: string): { ids: Set<string>; found: boolean } { | ||
| const target = parentId.trim().toLowerCase(); | ||
| const anchor = corpus.find((item) => item.id.trim().toLowerCase() === target); | ||
| if (!anchor) { | ||
| return { ids: new Set<string>(), found: false }; | ||
| } | ||
| const childrenByParent = buildChildrenByParent(corpus); | ||
| const descendants = collectDescendants(anchor.id, childrenByParent); | ||
| const ids = new Set<string>([anchor.id, ...descendants.map((item) => item.id)]); | ||
| return { ids, found: true }; | ||
| } |
There was a problem hiding this comment.
To ensure robust case-insensitive matching and prevent potential bugs where mixed-case IDs in the corpus fail to match, the ids set returned by collectSubtreeIds should store normalized (lowercase and trimmed) IDs. This guarantees that case-sensitive .has() checks on the resulting set work reliably regardless of the original casing in the files.
function collectSubtreeIds(corpus: ItemFrontMatter[], parentId: string): { ids: Set<string>; found: boolean } {
const target = parentId.trim().toLowerCase();
const anchor = corpus.find((item) => item.id.trim().toLowerCase() === target);
if (!anchor) {
return { ids: new Set<string>(), found: false };
}
const childrenByParent = buildChildrenByParent(corpus);
const descendants = collectDescendants(anchor.id, childrenByParent);
const ids = new Set<string>([
anchor.id.trim().toLowerCase(),
...descendants.map((item) => item.id.trim().toLowerCase())
]);
return { ids, found: true };
}| listedFrontMatter = listedFrontMatter.filter((item) => subtreeIds?.has(item.id)); | ||
| allItems = allItems.filter((item) => subtreeIds?.has(item.id)); |
There was a problem hiding this comment.
Since subtreeIds now stores normalized lowercase IDs, we must normalize the lookup keys to lowercase as well to ensure the case-sensitive Set.has() check succeeds.
| listedFrontMatter = listedFrontMatter.filter((item) => subtreeIds?.has(item.id)); | |
| allItems = allItems.filter((item) => subtreeIds?.has(item.id)); | |
| listedFrontMatter = listedFrontMatter.filter((item) => subtreeIds?.has(item.id.trim().toLowerCase())); | |
| allItems = allItems.filter((item) => subtreeIds?.has(item.id.trim().toLowerCase())); |
| const scopedAgenda = | ||
| subtreeIds === undefined | ||
| ? agenda.events | ||
| : agenda.events.filter((event) => subtreeIds.has(event.item_id)); |
There was a problem hiding this comment.
Normalize the event.item_id to lowercase when checking against subtreeIds to ensure case-insensitive matching for agenda events.
| const scopedAgenda = | |
| subtreeIds === undefined | |
| ? agenda.events | |
| : agenda.events.filter((event) => subtreeIds.has(event.item_id)); | |
| const scopedAgenda = | |
| subtreeIds === undefined | |
| ? agenda.events | |
| : agenda.events.filter((event) => subtreeIds.has(event.item_id.trim().toLowerCase())); |
…o-end (Gemini) collectSubtreeIds now stores trimmed-lowercase ids and the listed/all/agenda membership checks lowercase their lookup keys, making --parent subtree matching explicitly case-insensitive regardless of id casing in the corpus. Behaviour is unchanged for the canonical lowercase ids pm emits; this just removes any latent case-sensitivity assumption. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
|
@gemini-code-assist follow-up addressed in 63c70a2 + f9f9a54:
@coderabbitai review |
|
Kicking off a fresh review of the updated branch now. ✅ Action performedReview finished.
|
There was a problem hiding this comment.
Code Review
This pull request introduces several context-retrieval ergonomics improvements, including the --depth full option for pm context to provide a comprehensive snapshot of all sections without per-section row caps, and the --parent <id> option to scope the context snapshot to a single epic or parent subtree. Additionally, it adds the --no-truncate (alias --all) flag to pm list commands to override any limits and return all matched rows, while also reporting the pre-pagination total when results are truncated. Review feedback highlights a potential TypeError when filtering calendar events with null or undefined item_id values, and suggests slicing the ranked focus items by the specified limit when scoping by parent to ensure user-specified limits are respected.
Important
The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.
| const scopedAgenda = | ||
| subtreeIds === undefined | ||
| ? agenda.events | ||
| : agenda.events.filter((event) => subtreeIds.has(event.item_id.trim().toLowerCase())); |
There was a problem hiding this comment.
If event.item_id is null or undefined (which can happen for external or generic calendar events), calling event.item_id.trim() will throw a TypeError and crash the command. We should add a safety guard to ensure event.item_id is defined before calling .trim().
| : agenda.events.filter((event) => subtreeIds.has(event.item_id.trim().toLowerCase())); | |
| : agenda.events.filter((event) => event.item_id && subtreeIds.has(event.item_id.trim().toLowerCase())); |
There was a problem hiding this comment.
Declining — false positive. CalendarRow.item_id is a non-optional string (calendar.ts:66), always set from item.id when each row is built (calendar.ts:599/622/667) and consumed unguarded elsewhere (calendar.ts:851, 932). It cannot be null/undefined, so .trim() is safe here.
| allItems = allItems.filter((item) => subtreeIds?.has(item.id.trim().toLowerCase())); | ||
| } | ||
|
|
||
| const ranked = [...listedFrontMatter].sort((left, right) => compareCriticalItems(left, right, statusRegistry)); |
There was a problem hiding this comment.
When parentScope is defined, listedFrontMatter is fetched without a limit. If the high_level and low_level focus sections are not sliced by limit during their construction from ranked, they will display all active items in the subtree, overriding any user-specified --limit. To guarantee that the limit is respected, we should slice ranked by limit when parentScope is defined.
| const ranked = [...listedFrontMatter].sort((left, right) => compareCriticalItems(left, right, statusRegistry)); | |
| const sortedRanked = [...listedFrontMatter].sort((left, right) => compareCriticalItems(left, right, statusRegistry)); | |
| const ranked = parentScope === undefined ? sortedRanked : sortedRanked.slice(0, limit); |
There was a problem hiding this comment.
Declining — false positive. --limit is already respected: highLevel and lowLevel each .slice(0, limit) (lines 1228/1232), and agendaEvents slices too. Verified: pm context --parent <epic> --limit 1 returns 2 focus rows (1 high + 1 low), not the whole subtree. The unpaginated listedFrontMatter only ensures correct subtree resolution before the per-section slice.
…drift - pm-ldr1: follow-up Feature for the remaining GH-157 asks not covered by #238 (structured inline tag:/status: query syntax + matched-text highlighting). - pm history-repair pm-j0vc: re-anchor a pre-existing single-item hash mismatch (from #234) so `pm health` is ok:true again. Audited, verify_ok:true. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
|
PM context: pm-0c0j. Historical linkage verified during the 2026-08-04 ecosystem review. This comment records canonical context only; it does not reopen or change this artifact. |
Summary
Context-retrieval ergonomics that serve the core philosophy project management = context management — surface richer project context in fewer agent round-trips.
pm context(GH-156 / pm-j0vc, pm-ds0m)--depth full— comprehensive snapshot: every known section (ignoring the per-section settings toggles that brief/standard/deep respect) with no per-section row cap (overridable with an explicit--limit). New depth level inCONTEXT_DEPTH_VALUES.--parent <id>— scope the focus items, hierarchy, agenda, and all derived sections to the anchor item plus its transitive descendants — the "what is the status of this epic?" view for large trackers. Reuses the existingbuildChildrenByParent/collectDescendantshelpers (no duplicate tree walk); case-insensitive anchor match; unknown anchor →NOT_FOUND.filters.parentis echoed and the markdown header gains ascope: subtree of <id>line.pm list(GH-154 / pm-0c0j)--no-truncate(alias--all) — force the full matched set, overriding any--limitin one call.result.total— pre-pagination match count, emitted only when--limit/--offsetactually dropped rows (both compact and verbose paths), so agents know how many remain.Verified already-shipped + documented
pm getchild rollup for Milestone/Epic (children: {count, active, by_status}), helpful calendar warning, andpm aggregate --parentall verified working (shipped 4014ea6, unreleased). Closed.--include-bodyalready pulls item bodies in bulk in a single call; compact JSON stays token-light by default. Documented prominently inCOMMANDS.md/AGENT_GUIDE.md(avoids body-by-default, which would break the token-efficiency philosophy).--fieldsrow-level projection from GH-156 is split to a focused follow-up (pm-hnjf) since it requires reshapingContextResultfocus-row types;--sectionalready covers section-level customization.Wiring
types · register-list-query options ·
normalize{List,Context}Options· CONTEXT/LIST flag contracts + commander-string-option contracts · MCPpm_contextdescription · zsh/fish completions · contracts golden fixture · docs.Tests
Unit + integration tests for every new branch (depth full/limit, subtree found/leaf/missing, no-truncate/total/compact-path, normalize mappings). Full coverage gate green (97.7/91.6/97.7/98), static-quality gate, contracts check, and docs-links gate all pass.
Manual verification
Sandbox tracker:
context --depth full(10 sections, no cap),context --parent(subtree-scoped focus +scope:markdown line +NOT_FOUNDon bad anchor),list --no-truncate/--all(override +no_truncateecho),list --limit(totalsurfaced), and all existing commands regression-clean.Closes GH-156, GH-154, GH-155, GH-223.
🤖 Generated with Claude Code