Repository navigation
Conversation
The drain lists these four as ineligible because they carry no remedy, so no lane can take them and the board cannot say what would close them. Each now carries one, checked against the tree at origin/main. The forge allocator and the title ceiling keep their open rulings (J12, and the ceiling's unit and value): their remedies open "Waits on" and name the fix for each answer, so the drain still hands them back. The reach baseline and the lint docs JSON shape need no ruling. Refs: iss-2608210737264758, iss-2609020721396212, iss-2609252211487887, iss-2610100649471088 Assisted-by: Claude:claude-opus-5-5
Each intent's Audit Notes now carries its review verdict, ingested from an intent-auditor run on origin/main: - itd-3: 2 met, 2 met with concerns, 2 not met (ac-5, ac-6) - itd-6: 3 met, 1 inconclusive (ac-1) - itd-24: 3 met, 3 met with concerns, 2 not met (ac-2, ac-7) The ingest captured one issue per intent for the check still owed. Two defects the run met are captured too: reserve refuses a shipped intent, so an owed review cannot be reserved, and the auditor's output rule describes a fail-closed fence the file ingest does not have. Refs: iss-2610241427397332, iss-2610241426406882, iss-2610241431221889, iss-2610241430042742, iss-2610241430326301 Assisted-by: Claude:claude-opus-5-5
…65 and itd-2609151138388536 Each intent's Audit Notes now carries its review verdict, ingested from an intent-auditor run on origin/main: - itd-201: 3 met, 7 met with concerns; conditions 2 narrowed, 1 untested - itd-2609091034175565: 52 met, 1 met with concerns; conditions 9 survived, 3 narrowed, 1 untested - itd-2609151138388536: 6 met, 2 met with concerns; 4 conditions survived No criterion is not met or inconclusive, so no check is left owed. Assisted-by: Claude:claude-opus-5-5
Each intent's Audit Notes now carries its review verdict, ingested from an intent-auditor run on origin/main: - itd-2609212103568351 (the status block): 1 met, 4 met with concerns - itd-2609212103572513 (release targets): 3 met, 1 met with concerns - itd-2610030720038073 (the home folder's name): 3 met, 2 met with concerns, 1 not met (ac-6: the indexing receipt measured a probe folder outside the home, and the run-log check was never taken) The ingest captured iss-2610241551259104 for the check ac-6 leaves owed. Refs: iss-2610241551259104 Assisted-by: Claude:claude-opus-5-5
…d-2610090831227812 Each intent's Audit Notes now carries its review verdict, ingested from an intent-auditor run on origin/main: - itd-2610030810350727 (the question layout): 4 met, 6 met with concerns, 1 inconclusive (ac-6, a criterion the intent struck) - itd-2610090831227812 (the person's command menu): 6 met, 6 met with concerns The ingest captured iss-2610241604049754 for the check ac-6 leaves owed. Refs: iss-2610241604049754 Assisted-by: Claude:claude-opus-5-5
…on sweep The triage readers confirmed defects outside the records they were triaging; each is filed with a remedy, one record per finding (the twelve duplicated decision records, found by four batches, are one). Also filed: the SessionEnd hook the host still cancels on a loaded machine, reported by the person with the remedy they chose, and the plugin catalogue's description, which differs from the manifest's. The coordination sweep record is resolved: the research note committed in 019a6c5 answers its four questions, and the intent it was owed to has since been superseded by a shipped one. Resolves: iss-2608230943533581 Refs: iss-2610241519088123, iss-2610241549052683, iss-2610241549107888, iss-2610241549164432, iss-2610241549220956, iss-2610241549275970, iss-2610241549334990, iss-2610241549396753, iss-2610241549433260, iss-2610241549487283, iss-2610241549531180, iss-2610241549584964, iss-2610241550026188, iss-2610241550073708, iss-2610241550119108, iss-2610241550154898, iss-2610241550202749, iss-2610241550257430, iss-2610241550303964, iss-2610241606415232 Assisted-by: Claude:claude-opus-5-5
The intent's Audit Notes now carries its review verdict, ingested from an intent-auditor run on origin/main: 4 met, 8 met with concerns, 1 not met (ac-13: the live walk's receipt was never attached to the pull request). The ingest captured iss-2610241613191748 for the check ac-13 leaves owed. Refs: iss-2610241613191748 Assisted-by: Claude:claude-opus-5-5
Each fix is the smallest edit that makes a page match the code: - install: the name-guard hooks run abcd only where a clone opts in, and the Stop, StopFailure and SubagentStop hooks download nothing either - upgrade to v0.14.0: the launch preview exits 0 whatever it finds and fails only when it cannot run - roles: a fresh install starts with a person as facilitator (adr-2610150603484802), and the register setting is a design target - terminology: the implement loop is abcd's own, the pacing intent and an MCP server ship, and itd-33 is superseded by the shipped reservation - writing-style: name the persona_registry and retired-role-word rules - reference index: list the pages the folder actually holds Assisted-by: Claude:claude-opus-5-5
…6-10-16 Each intent's Audit Notes now carries its review verdict, ingested from an intent-auditor run on origin/main: - itd-2610161016391855 (a stuck drain lane hands itself back): 8 met, 3 met with concerns - itd-2610161016394967 (several people's drains share the work): 13 met, 7 met with concerns - itd-2610161016401115 (a drain runs unattended on a hosted session): 6 met, 5 met with concerns - itd-2610161616086564 (after a plugin update abcd offers setup): 7 met, 3 met with concerns No criterion is not met or inconclusive, so no check is left owed. Assisted-by: Claude:claude-opus-5-5
The docs-currency review of docs/ and the root README checked 558 claims. Sixteen were fixed on this branch. These 47 records hold the rest: - 30 from the install, how-to, explanation and reference pages, and from the first two parts of the generated CLI reference (readers D1 to D5) - 12 from the last two parts of the CLI reference (readers D6 and D7). The page is generated from the Go help text, so each record names the help string to change rather than the page. - 5 from the root README (reader D8). Another session is editing that file, so its findings are recorded here, not edited. One more README finding is already held by an open record and was not filed again. Security-class claims were counted and not reviewed. Assisted-by: Claude:claude-opus-5-5
The intent's Audit Notes now carry its review verdict, ingested from an intent-auditor run on origin/main. After a plugin update, abcd says what changed and offers the setup the release brings: 4 criteria met and 4 met with concerns. Of its four scope conditions, two survived and two narrowed. No criterion is not met or inconclusive, so no check is left owed. That pays the last of the reviews this drain could take: 17 of the 21 owed are recorded, and the other 4 are left for the session that handles security-class work. Assisted-by: Claude:claude-opus-5-5
One landing for three record-and-docs branches from the 2026-10-24 low-CPU run, so the crowded machine runs one preflight instead of three: - triage/handback-records: the hand-back triage's resolves and remedies - audit/owed-fidelity-reviews: 17 owed fidelity reviews recorded - docs/currency-2026-10-24: the docs-currency fixes and their captures Assisted-by: Claude:claude-opus-5-5
Thirteen read-only readers classified every section of the brief against the files that state the same thing: the command pages, docs/, the ADRs, the principles, the intents and the code. Where a section conflicted with its counterpart or with the code, or described machinery that has since moved, the reader recorded a finding with the edit that fixes it. These 105 records are those findings, one each. Most are a second copy of a verb's behaviour that has drifted from the command page or the code, which is the study's main result: the surface chapters and commands/ both state each verb's behaviour, and only the chapter is gated. Eighteen findings were already held by open records and were not filed again; one more turned out to duplicate an open record when it was filed, and was dropped. Where the filing-time match linked a record to an unrelated one, the link was removed; the seven links kept were checked. Sections touching security were counted and not assessed. Assisted-by: Claude:claude-opus-5-5
…eft open Two sets of findings from this run that had no record yet. Six came from the state-of-the-art passes run during the hand-back triage, outside the records being triaged: two stale counts in the rulesets README and the CI header, the kernel-format claim for the Assisted-by trailer, a dated statement in a research note, a missing link gate for the docs site, and the cloud connector's undocumented-answer comment. Fifteen came from the fidelity reviews' gap audits: promises the delivery left unbuilt or unproven, each with the step that closes it. Three of the reviews met the same unrecorded lab rerun, which is one record. One is a defect in the audit itself: it judges acceptance criteria a later decision struck, so the owed check such a criterion files can never clear. Missing items an open record already holds were not filed again. Assisted-by: Claude:claude-opus-5-5
The brief's single-source study found both asks already met on main: - The memory-note glossary entry names /abcd:memory add as shipped. The record said no such verb existed; c8d1ab0 shipped it, so the entry is right as written. - The configuration chapter's user-scope inventory now lists interviews/ and backups/, added in a394eb1, which is what that record asked for. The other omissions the study found are held by an open record. Resolves: iss-2610141705112796, iss-2610161332060895 Assisted-by: Claude:claude-opus-5-5
Two records quoted text that the renderer refuses outside a code span: a README link to CITATION.cff, which does not resolve from the ledger, and two paths with a placeholder in angle brackets, which read as inline HTML. Both quotes now sit in code spans, so record-lint passes them and the quoted text is unchanged. Assisted-by: Claude:claude-opus-5-5
REPPL
enabled auto-merge
October 26, 2026 10:27
github-merge-queue
Bot
removed this pull request from the merge queue due to a conflict with the base branch
Oct 26, 2026
Main moved while the landing waited in the queue. One conflict, in docs/reference/terminology.md: this branch rewrote the Agent harness and Agent loop rows, and main rewrote the Agent identity row. Each side's rows are kept. Assisted-by: Claude:claude-opus-5-5
REPPL
enabled auto-merge
October 26, 2026 14:54
This branch has not been deployed
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
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.
Records and docs from the 2026-10-24/25 low-CPU run, landed as one branch so the shared machine runs one preflight instead of four. The branch adds no Go code: it carries issue records, intent audit notes and docs/ prose.
What it lands
Not in this pull request
Checks
make preflightpassed on this branch's tip, run through the load gate.Assisted-by: Claude:claude-opus-5-5