Skip to content

[Bug]: Greenfield T-002 seed still fails the inception schema gate after T-3948 (fw init path not covered) #86

Open
Open
[Bug]: Greenfield T-002 seed still fails the inception schema gate after T-3948 (fw init path not covered)#86
Labels
area:initfw init / greenfield seedsbugSomething isn't workingseverity:mediumWrong behaviour with a workaroundstatus:releasedReleasedtriagedClassified by the AEF intake pipeline

Description

@MikeEchoVoid
MikeEchoVoid
opened on Oct 7, 2026

Bug Description

T-3948 (d20c996) fixed the deadlock on the inception schema gate for upgraded consumers: the gate now judges the post-edit file, and fw upgrade backfills target_blast_radius / voi_score. The fw init path was not covered: the greenfield seed for T-002 still ships without the two fields, so a freshly initialised project starts with an inception that fails the gate.

Evidence, 1.8.6 (1ea1044):

  • lib/seeds/tasks/greenfield/T-002-define-project-goals.md: workflow_type: inception, no target_blast_radius, no voi_score. Last touched by 89bed4f (T-3636), before T-3948.
  • lib/init.sh copies the seed (line 714) and never calls lib/inception_schema_backfill.py; only lib/upgrade.sh does (line 1741).
  • tests/unit/t3948_inception_schema_upgrade.bats covers "fw upgrade of an initialised consumer backfills its old inception"; none of the eight tests in that file covers a project that is initialised and never upgraded (I did not search the other test files).

Steps to Reproduce

  1. fw init a greenfield project with provider claude.
  2. Open a Claude Code session in it and run fw work-on T-002.
  3. Ask the agent to tick an Agent AC or to fill in ## Recommendation in the T-002 task file.

Observed end to end on 1.8.3. On 1.8.6 this is read from the source; I have not run a fresh fw init.

Expected Behavior

  1. fw init a greenfield project with provider claude.
  2. Open a Claude Code session in it and run fw work-on T-002.
  3. Ask the agent to tick an Agent AC or to fill in ## Recommendation in the T-002 task file.

Observed end to end on 1.8.3. On 1.8.6 this is read from the source; I have not run a fresh fw init.

Actual Behavior

  1. fw init a greenfield project with provider claude.
  2. Open a Claude Code session in it and run fw work-on T-002.
  3. Ask the agent to tick an Agent AC or to fill in ## Recommendation in the T-002 task file.

Observed end to end on 1.8.3. On 1.8.6 this is read from the source; I have not run a fresh fw init.

fw doctor output



AI Agent


Claude Code


Operating System


Linux


Additional Context


Version: 1.8.6 (1ea1044), vendored consumer. First seen on 1.8.3 (542623a).


Suggestions:


    

  1. Add the two fields to the greenfield T-002 seed, or run the T-3948 backfill at the end of fw init.
    2. 

  2. Add a check (test or fw doctor) that every seeded task passes the gates that apply to its workflow_type. T-2862 and T-3948 were both deadlocks on this same seeded task.
    3. 

  3. In the block message, say that FW_ALLOW_INCEPTION_SCHEMA_DRIFT=1 is an operator action when the blocked call is a Write/Edit tool call: the agent cannot set an environment variable for a hook.
    4. 

Activity
  1. MikeEchoVoid
    added 
    bugSomething isn't working
     on Oct 7, 2026
  2. DimitriGeelen commented on Oct 10, 2026 
    @DimitriGeelen
    DimitriGeelen
    on Oct 10, 2026
    Owner
    [AEF intake] Received — 2026-10-10
    

    Thank you. This issue has been ingested into the Agentic Engineering Framework's intake as INTAKE-0002. Next step: classification (type, affected area, severity, whether it reproduces on the current release and on bleeding-edge). You will get a comment here at each step: classified → decision → in progress → fixed on bleeding-edge → available for testing → released.
    

    These status updates are automated, but Dimitri personally looks at every issue and decides what happens to it. Replies on this issue are read.
  3. DimitriGeelen commented on Oct 10, 2026 
    @DimitriGeelen
    DimitriGeelen
    on Oct 10, 2026
    Owner
    [AEF intake] Classified — 2026-10-10
    

      

    • Type: bug
      • 

    • Area: fw init greenfield seeds / inception schema gate
      • 

    • Severity: medium
      • 

    • Status: confirmed from the source as still present in the current release (v1.8.8) and on bleeding-edge
      • 

    

    [AEF intake] Accepted — 2026-10-10
    

    Tracked as T-4040, prioritised (horizon: now). Next update here: in progress.
    

    These status updates are automated, but Dimitri personally looks at every issue and decides what happens to it. Replies on this issue are read.
  4. DimitriGeelen
    added 
    triagedClassified by the AEF intake pipeline
     
    status:acceptedAccepted; linked to an AEF task
     
    severity:mediumWrong behaviour with a workaround
     
    area:initfw init / greenfield seeds
     on Oct 10, 2026
  5. DimitriGeelen
    added a commit that references this issue on Oct 11, 2026
    T-4039: inception — inbound GitHub issue lifecycle (ingest, acknowled…
    f633d9c
  6. DimitriGeelen commented on Oct 11, 2026 
    @DimitriGeelen
    DimitriGeelen
    on Oct 11, 2026
    Owner
    [AEF intake] In progress — 2026-10-11
    

    Work has started under T-4040. Approach: the greenfield T-002 seed gets the two fields the inception schema gate requires, and a new test checks that every seeded task passes the gates for its type (your suggestion 2), so this class cannot recur silently. Next update here: fixed on bleeding-edge.
    

    These status updates are automated, but Dimitri personally looks at every issue and decides what happens to it. Replies on this issue are read.
  7. DimitriGeelen
    added 2 commits that reference this issue on Oct 11, 2026
    T-4040: real ACs for GitHub #86 (seed fields, every-seed-passes-its-g…
    72274ac
    T-4040: greenfield T-002 seed carries target_blast_radius/voi_score; …
    20289bf
  8. DimitriGeelen commented on Oct 11, 2026 
    @DimitriGeelen
    DimitriGeelen
    on Oct 11, 2026
    Owner
    [AEF intake] Fixed on bleeding-edge — 2026-10-11
    

    Fixed under T-4040 in commit 20289bf (on the bleeding-edge branch; not yet in a release).
    

      

    • The greenfield seed task T-002 that fw init creates now carries target_blast_radius and voi_score, so the inception schema gate no longer blocks edits to it.
      • 

    • A new test checks every shipped seed task against the gates that guard it, so a seed cannot fall behind a schema change again.
      • 

    • The gate's message now names its bypass as an operator action.
      • 

    

    For a project already initialised: add target_blast_radius: 3 and voi_score: 0.5 to T-002's frontmatter (the values the seed now ships). To test the fix on a fresh init: git clone -b bleeding-edge https://github.com/DimitriGeelen/agentic-engineering-framework.git aef-be && aef-be/bin/fw init /path/to/new-project. Feedback here is welcome. Next update here: released.
    

    These status updates are automated, but Dimitri personally looks at every issue and decides what happens to it. Replies on this issue are read.
  9. DimitriGeelen
    added a commit that references this issue on Oct 11, 2026
    T-4039: #86/#87 Fixed-on-bleeding-edge posted (intake ledger); file T…
    41a7d70
  10. DimitriGeelen
    added a commit that references this issue on Oct 18, 2026
    T-4339: retrospective of the 22 external GitHub issues (drafts only; …
    8e86568
  11. DimitriGeelen commented on Oct 18, 2026 
    @DimitriGeelen
    DimitriGeelen
    on Oct 18, 2026
    Owner
    

    [AEF intake] Released — 2026-10-18
    

    T-4040 is released in v1.8.11.
    

    Please confirm it fixes this for you. Reply here if it does not. With no reply in 14 days this issue is closed with a note, and you can reopen it.
    

    How it was tested: tests/unit/t4040_seeds_pass_schema_gate.bats (2 tests: every inception seed passes the schema gate when written as a task; the seed tree is non-empty, so the check is not vacuous). Green on bleeding-edge.
    
Why it is released now: v1.8.11 passed the full-suite comparison (https://github.com/DimitriGeelen/agentic-engineering-framework/blob/50fae848ff46e156c1be630e85b0644aa070be38/docs/reports/T-4246-full-suite-comparison.md), the operator's release decision, and the field-test record for the release.
    
Field evidence: none yet that names this fix. Please confirm on your greenfield run.
    
Release: https://github.com/DimitriGeelen/agentic-engineering-framework/releases/tag/v1.8.11 · Field report: https://github.com/DimitriGeelen/agentic-engineering-framework/blob/50fae848ff46e156c1be630e85b0644aa070be38/docs/reports/field-test-v1.8.11-results.md
    

    These status updates are automated, but Dimitri personally looks at every issue and decides what happens to it. Replies on this issue are read.
  12. DimitriGeelen
    added 
    status:releasedReleased
     on Oct 18, 2026
  13. DimitriGeelen
    added a commit that references this issue on Oct 19, 2026
    T-4339: posting phase recorded — Released on #86-#94, Received/Classi…
    cfcea4b
  14. DimitriGeelen
    removed 
    status:acceptedAccepted; linked to an AEF task
     on Oct 19, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Metadata
Metadata
Assignees
No one assigned
    Labels
    area:initfw init / greenfield seedsbugSomething isn't workingseverity:mediumWrong behaviour with a workaroundstatus:releasedReleasedtriagedClassified by the AEF intake pipeline
    Projects
    No projects
      Milestone
      No milestone
      Relationships
      None yet
      Development
      No branches or pull requests
      Issue actions