docs(research-patches): capability-census — full-population enumeration (Superset/aif/CC-native, agnostic lens) - #329
Merged
Conversation
…on of Superset/aif-handoff/CC-native (agnostic lens) Closes the T10 population-enumeration gap the survey (#321) + harvest (#328) left open: each product's full capability population was never enumerated. Two dual-channel-verified criterion-zero findings hitting live problems — F1 CC paths: frontmatter <-> our inject-matching-rule.sh (ADAPT dual-channel, 'как с хуком'); F2 CC Routines /fire <-> broken runtime-bridge dispatch (REFERENCE). SSOT rows #101-#103 appended (#100 reserved for #328). Prior-art: skipped — research patch + SSOT append-only rows #101-#103, no new capability/code/dependency (T5 research-only)
…p (Superset/Superpowers/AI-Factory/amux/CC-native) Autonomous parallel Mode-A sweep (4 Opus workers) per maintainer 'пройтись по всем спутникам'. OhMyOpencode + Cursor excluded per maintainer. §10 added + SSOT #104-#108. Spot-check caught propagated error: '#85=amux' is wrong (#85 = pre-commit); amux had no row -> #104 + correction. Top payload: CC-native under-used features (/goal, orchestrator-gate hooks, /code-review --fix). CC claims worker-sourced, flagged for bench-test before live wiring. Prior-art: skipped — research patch §10 extension + SSOT append-only #104-#108, no new capability/code/dependency (T5 research-only)
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Closes the T10 population-enumeration gap the companion survey (#321) and own-stack harvest (#328) left open — each product's full capability population was never enumerated (incident-driven accretion, not exhaustive sweep). Maintainer-directed 2026-06-01: «ищи и записывай … агностик — CC когда есть, своё когда нет, как с хуком».
Population quantified: Superset SSOT had 3 rows for ~30 capabilities; CC knew ~5 hook-events vs ~29 documented; aif ~13 point-rows vs ~7 packages + ~20 env-vars. Full census in patch §2.
Two dual-channel-verified criterion-zero findings hitting LIVE problems:
paths:frontmatter ↔ ourinject-matching-rule.sh. CC natively path-scopes rules (verified/en/memory+/en/hooks). Already partially dogfooded in-repo (phase-research-coverage.md usespaths:). Verdict ADAPT dual-channel per the maintainer principle «как с хуком»: nativepaths:on CC + our hook as portable fallback for non-CC consumers; rules carry both markers./fire) ↔ broken runtime-bridge dispatch. Working HTTP-dispatch substrate (verified/en/routines). Verdict REFERENCE shipped / ADOPT-when-needed operator.SSOT: +3 append-only rows #101–#103 (#100 reserved for #328; no existing row edited). EXTENDS #328 + #321, supersedes neither.
2 DECISION-NEEDED surfaced (reviewer-discipline §2, not decided) — F1/F2 each touch live code → own follow-up R-phase.
Patch:
docs/meta-factory/research-patches/2026-06-01-capability-census.md(188 lines).§1.7 Forward-check applied
Complies with build-first-reuse-default §1.1 (own-stack-first ran first; F1/F2 are CC-native-supersedes findings) per
.claude/rules/build-first-reuse-default.md:35; dual-implementation-discipline §3 (the F1 verdict IS dual-channel: CC-native + portable fallback) per.claude/rules/dual-implementation-discipline.md; no-paid-llm §2 (F3 subscription-bundled; research = WebFetch+DeepWiki) per.claude/rules/no-paid-llm-in-ci.md:17; doc-authority (<!-- scope:capability-census -->first line + header) per.claude/rules/doc-authority-hierarchy.md:1; reviewer-discipline §2 (F1/F2 surfaced not decided) per.claude/rules/reviewer-discipline.md:1.§1.7 Backward-check applied
EXTENDS
2026-06-01-satellite-feature-harvest.md(#328) +2026-06-01-companion-capability-survey.mdby closing their T10 gap — both cited, neither superseded. SSOT #100 reserved for #328 (no collision); rows #28/#29/#30/#46/#65/#85/#86/#88/#97/#98/#99 cited, none edited; #101-#103 append-only. No shipped artefact retired —inject-matching-rule.shKEPT as portable fallback (F1),questions.tsuntouched. Append-only confirmed:git diff= +191 insertions, 0 deletions.