/ndd rules what it can articulate, and reports at the finding's length - #309
Conversation
Closes #307 #297 gave /nr "if you can articulate it, send it back — not forward" and #298 carried it to /nm and /ncc. /ndd was the last read still ending "This command does not rule", so all three of its findings produced the same full presentation. Two of the three are answers this command can state: an already-answered question exits with the existing answer as the work order, a misroute with the correction. Both are ruled here through human_rule, and both report in a few lines. The genuine question is unchanged — the human's judgement, the full presentation. The report shrinks, the read does not: which of the three findings is in hand is only knowable after deriving the issue's intent, reading the diff against it and testing the question's premise. The row's url opens every form of the report, the short one included (#304). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
|
Warning Review limit reached
Next review available in: 11 minutes Limit details: You’ve used all 1 included review currently available under your plan. You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. How can I continue?After more reviews become available, a review can be triggered using the To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews. How do review limits work?CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability. For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window. Please refer docs for additional details. Review details⚙️ Run configurationConfiguration used: Organization UI Review profile: ASSERTIVE Plan: Pro Plus Run ID: 📒 Files selected for processing (5)
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 |
1 participant
Closes #307
/nddwas the last read command still ending "This command does not rule — itis the read that precedes the human's word", so all three of its findings
produced the same full presentation. #297 replaced that in
/nrwith if youcan articulate it, send it back — not forward and #298 carried it to
/nmand/ncc; this uses the same wording and the same section shape rather than athird phrasing.
What changed
/nddrules what it can articulate.human_rulejoins the grant, and twoof the three findings exit here: an already-answered question is ruled with
the existing answer as the work order and the pointer in the note; a misroute
is ruled as the correction. What goes forward is the answer this command does
not have — a genuine judgement call, and an already-answered question whose
pointer is an unverifiable recollection.
question: the full presentation, unchanged. The other two: the pointer or the
correction, the ruling taken, the work order — a few lines.
finding is only reachable by deriving the issue's intent, reading the diff
against it and testing the question's premise, and which of the three is in
hand is knowable only afterwards.
urlopens every form of the report (/ndd reports omit the PR url, so the reader has to reconstruct the link #304), the one-line formincluded, as its own requirement rather than an instance of "print every
field".
ndd_grants_the_queue_the_pr_the_source_the_lens_and_the_send_backinpr-review-report-rs/src/main.rs./nddalready had the equivalent of thetest that caught
/nr's grant change in /nr sends back anything it can articulate, instead of presenting it #297, and it should keep it: it iswhat stops a later edit leaving prose that promises a send-back with no tool
to make one.
Beyond the issue
Two statements went false when a read command was first granted a write and
would have gone falser here, so they are corrected in the same diff:
plugins/human-fsm/README.md— the/nrand/nddrows both said "Writes noGitHub state", and neither does now.
.claude-plugin/marketplace.json— "none writes GitHub state itself".QA
ndd_grants_the_queue_the_pr_the_source_the_lens_and_the_send_back— fails on base (git checkout origin/main -- plugins/human-fsm/commands/ndd.md, thencargo test --bin pr-review-report marketplace_tests: left has four MCP tools, right has five,human_rulemissing)ndd.mdfrontmatterallowed-tools→ dropmcp__plugin_human-fsm_fsm__human_rule(the base-file revert above IS that mutation) → killed by that test. Baseline and post-change full suite both green: 1329 + 127 tests, 0 failedhuman_ruleas /nm and /ncc send back what they can articulate #298 grants it to/nm; three report lengths; keep the row'surlper /ndd reports omit the PR url, so the reader has to reconstruct the link #304) and the settled wording in the /nr sends back anything it can articulate, instead of presenting it #297 and /nm and /ncc send back what they can articulate #298 diffs, not the existing/nddtexthuman_rule, (b) report length per finding stated in the command, (c)urlin every case — all three covered; the read-must-not-shrink guard is stated as its own paragraph. /ndd reports omit the PR url, so the reader has to reconstruct the link #304 isRefs, not closed: this PR carries theurlrequirement for/nddonly, and /ndd reports omit the PR url, so the reader has to reconstruct the link #304 also asks for/nr,/ncc,/nmand a shipped-command test asserting it🤖 Generated with Claude Code