Part D: live-trigger mutation-hunt, confirm the pin bump, find a new blocker - #201
Open
wesleyscholl wants to merge 3 commits into
Open
Part D: live-trigger mutation-hunt, confirm the pin bump, find a new blocker#201wesleyscholl wants to merge 3 commits into
wesleyscholl wants to merge 3 commits into
Conversation
…mp, find a new blocker Manually dispatched the mutation-hunt workflow_dispatch job against lopi-ratelimit (PF-0b's known-non-trivial target, 11 real surviving mutants of 51 tested) now that Part A's KIBAN_REF v1.14.0 bump is on main. Confirmed live, not assumed: the pin bump fixes what it set out to fix -- kiban v1.14.0 clones clean, bin/kiban-mutation-hunt exists and runs, coverage generates, the loop starts and identifies a real uncovered mutant target. Found a second, separate, real blocker while doing this: ANTHROPIC_API_KEY doesn't reach the job (tokens=0 cost=$0.0000, generation_failed). Not a code problem -- a GitHub Actions repository-secret question outside what this session can fix. Full round-by-round output and reasoning in LEDGER.md's Mutation-Hunt-Live-Trigger entry; carried forward as the next blocker in NEXT_SESSION_PROMPT.md. No version bump -- verification run, not a code change. Konjo-Sprint: mutation-hunt-live-trigger Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013jrS3y6XStkhXaBm4DCYDr
wesleyscholl
marked this pull request as ready for review
August 16, 2026 19:56
…live-trigger-d # Conflicts: # LEDGER.md # NEXT_SESSION_PROMPT.md
…ss re-verifies dry_check.py's --changed-only/--staged-only early-return path (no source files touched, a real and legitimate outcome for a docs-only PR) never wrote --report's output file, but konjo-gate.yml's DRY check step unconditionally reads it -- a docs-only PR (this one) is the first to hit it, crashing with FileNotFoundError instead of showing the true zero-violation result. Added _write_empty_report() so the early-return path writes a real count:0 report, verified locally against the exact CI invocation. Also re-verified 3 state docs whose 20-commit staleness cap this PR's own commit volume tripped (the recurring whack-a-mole this repo's own LEDGER.md already documents for this gate): - docs/ops/PANIC_AUDIT.md: re-ran the exact cited deny-flag clippy command workspace-wide, clean. - docs/security/EGRESS_SURFACE.md: confirmed its two cited handlers.rs line citations are unshifted and content-identical. - docs/LOOP_ENGINEERING_ROADMAP.md: found and fixed two real citation drifts from the Collision-Oracle-Build sprint's AgentRunner wiring (with_skills 92->94, seed_skills 210-241->230-244, record_skill_activation ->330-343) -- content unchanged, only line numbers corrected. Konjo-Sprint: mutation-hunt-live-trigger Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013jrS3y6XStkhXaBm4DCYDr
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Part D of the combined P3a-closeout/Collision-Oracle/RepoProfile-parity/mutation-hunt/branch-triage sprint. Both gating parts (A, B/C) are now merged, so this manually dispatches the real
mutation-huntworkflow_dispatchjob againstlopi-ratelimit— PF-0b's own known-non-trivial target (11 real surviving mutants of 51 tested).Verdict: the
KIBAN_REFv1.14.0 pin bump is confirmed fixed, live, not assumed. Two real dispatch runs:31950963613): failed instantly on my own mistake — passed a short SHA asdiff_base_ref, andgit fetch origin "$ref"needs a resolvable name.31951537127): re-dispatched with the full 40-char SHA. This time it ran for real — kibanv1.14.0clones clean,bin/kiban-mutation-huntexists and runs (the exact gapv1.8.0had), coverage generates, and the loop starts and finds a real uncovered mutant target.A second, separate, real blocker found along the way:
ANTHROPIC_API_KEYdoesn't reach this job — round 1 terminatedgeneration_failedattokens=0 cost=$0.0000. Checked before concluding: this repo's otherANTHROPIC_API_KEYconsumer (G5 · Adversarial Review) is designed to soft-fail-safe on a missing key, so its "success" on every prior run isn't evidence the secret is configured. Not fixed here — outside what this session can do (no repo-settings access) and shouldn't be guessed at. Full round-by-round output, the exact log lines, and reasoning inLEDGER.md'sMutation-Hunt-Live-Triggerentry; carried forward inNEXT_SESSION_PROMPT.md.No version bump — this is a verification run, not a code change.
Test plan
mutation-huntworkflow twice, confirmed via job logs, not assumedG5's soft-fail design before treating its past "success" as evidence the secret worksGenerated by Claude Code