Conversation
#1726, #1728) - scripts/statusline.py CHANNEL_STATES now maps supertool 0.64.0's sixth channel:health verdict, "BOUND, UNPROVEN" (bound, verified and subscribed but has never forwarded anything), to its own state and its own glyph (ch(unproven), ASCII chu) instead of falling through to cannot_determine (#1726). - agents/doctor.md now clears its own role marker as its last act before reporting, mirroring the sub-manager's tick_handback.py --clear-marker-root (#1585). A doctor spawn wrote its marker (#1690) but never cleared it, leaving every ordinary run's marker live for up to MARKER_TTL_SECONDS and making the very next tick's own sub-manager refuse to declare its role -- the #1716 conflict refusal, tripped by the doctor's own residue rather than a real rival (#1728). Both diffs raise the budgeted files they touch (agents/doctor.md, CLAUDE.md) per the repo's own editing-exception convention, with the weighed sentence recording it in each. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
…ing (#1726) Self-review (Explore + oss:auditor, both independently) found the same two gaps left by the #1726 fix: - scripts/statusline.py still had five stray "five states"/"five recognised" mentions after the diff updated only two of the module's seven such docstrings/comments -- one of them ("scaled down for a field with five possible states") sitting seven lines above the diff's own corrected "six upstream states" in the very same docstring. All five brought to six. - scripts/doctor_check_channel_health_agreement.py's _health_signal had no branch for the new "unproven" raw_state, so channel_health_agreement_state reported could-not-compare for a repo in BOUND, UNPROVEN even though a single, real consumer was identified -- the same single-consumer shape forwarding/not_delivering/not_subscribed already agree against. Added "unproven" to the single-consumer tuple, plus a red/green test. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
…how" substring (CI) PR #1733's CI failed 4/10 legs on tests/test_scaffold_cli.py:: test_show_includes_owned_files_even_when_every_template_already_exists, disjoint from this diff's own files. Root cause: scaffold.py's `--show` embeds the full content of .oss/statusline.py (an owned file this repo vendors from scripts/statusline.py) inline in its output, and the new BOUND-UNPROVEN glyph comment added in a19ea82 read '"just started, nothing to show yet"' -- the literal substring "nothing to show" the test asserts never appears anywhere in --show's output. Reworded to "no traffic yet", same meaning, no collision. Confirmed the test passes locally, alone and as part of the wider scaffold/statusline/doctor test files (282 passed). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
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
1 participant
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
scripts/statusline.py'sCHANNEL_STATESnow maps supertool 0.64.0's sixthchannel:healthverdict,BOUND, UNPROVEN(bound, verified and subscribed, but has never forwarded anything), to its own state (unproven) and its own glyph (ch◔, ASCIIchu) instead of falling through tocannot_determine. Every stale "five states" docstring/comment in the file updated to "six". Closes statusline CHANNEL_STATES has no key forBOUND, UNPROVEN, the sixth channel:health state supertool added in 0.64.0 #1726.scripts/doctor_check_channel_health_agreement.py's_health_signalgained a branch for the newunprovenraw_state (found in self-review: it fell through toNone/could-not-ask, reportingcould-not-compareeven though a single real consumer was identified).agents/doctor.mdnow clears its own role marker (agent_role.py --clear --root .) as its last act before reporting, mirroring the sub-manager's owntick_handback.py --clear-marker-root(The spawn guard refused the scheduler its own tick: a repo-global role marker cannot tell who is asking, and the remedy it prints is wrong for the caller it blocked #1585). Previously a doctor spawn wrote the marker (oss:doctor applies, commits to main, pushes and writes permissions despite an explicit no — needs a guard in code, not prose #1690) but never cleared it, leaving it live for up toMARKER_TTL_SECONDSand making the very next tick's own sub-manager refuse to declare its role -- the A doctor run's role marker can overwrite a live sub-manager marker, silently disabling that tick's release refusal #1716 conflict refusal, tripped by the doctor's own residue rather than a real rival. Closes doctor spawn never clears its role marker, so /oss:run's first tick after step 1 always refuses (#1716 composition) #1728.CLAUDE.md,scripts/agent_budgets.py,scripts/claude_md_budget.py: budget-table re-baselines for the grownagents/doctor.mdandCLAUDE.mdfiles, per this repo's own third editing exception.Review
Two parallel review spawns (Explore, oss:auditor) against the first commit both independently found: (a) 5 remaining stale "five states" mentions in
scripts/statusline.py, and (b) the_health_signalmapping gap above. Both fixed in a follow-up commit, including a new regression test (test_unproven_channel_health_is_a_single_consumer_reading_1726, confirmed red before / green after). A mandatory second-pass round (triggered by the fix commit touching 4 files) returned no further findings.Test plan
tests/test_statusline_channel_613.py-- red before (1 failed, assert 5==6) / green after (54 passed)tests/test_doctor_marker_clear_1728.py(new) -- red before (3 failed) / green after (3 passed)tests/test_doctor_channel_health_agreement_860.py-- red before (1 failed on the new test) / green after (31 passed)lane_setup.py --claim/--lane) -- all passed, 1 pre-existing unrelated skippytestsuite not run locally per this repo's own rule; CI's 13 legs are the real run🤖 Generated with Claude Code
[AI-generated]