Skip to content

feat: autoresearch optimisation of all 48 commands#265

Open
tractorjuice wants to merge 1 commit intomainfrom
feat/autoresearch-all-commands
Open

feat: autoresearch optimisation of all 48 commands#265
tractorjuice wants to merge 1 commit intomainfrom
feat/autoresearch-all-commands

Conversation

@tractorjuice
Copy link
Copy Markdown
Owner

@tractorjuice tractorjuice commented Mar 27, 2026

Summary

Rebased onto main (v4.6.12) on 2026-04-19. Ships as v4.7.0 to reflect the scope of the compliance uplift.

Content uplift (48 commands)

Deep autoresearch runs (iterative scoring, 2–16 iterations each):

Command Score Key Improvement
risk 8.0 → 9.6 Orange Book 2023 alignment — replaced retired 4Ts with six treatment options (Avoid/Take/Retain/Change Likelihood/Change Consequences/Share), Three Lines Model, tightened score boundaries
glossary 7.6 → 9.2 Project-contextualised definitions, Terminology Consistency Notes
sobc 8.0 → 9.4 Green Book 2026 alignment (Theory of Change, AST, VFM)
secure 8.0 → 9.6 CAF Maturity Summary, Security Remediation Roadmap, Principles-to-CAF Matrix
requirements 8.0 → 9.2 MoSCoW, NFR sub-prefixes (NFR-P/SEC/ACC/REL), Requirements Coverage Matrix
stakeholders 8.4 → 9.0 Impact Assessment, Change Readiness, Engagement Priority
service-assessment 8.0 → 8.8 Readiness Score, Evidence-to-Standard Traceability
tcop 8.0 → 9.0 Compliance Score, Principles-to-TCoP Traceability
dpia 8.0 → 9.0 DPIA Risk Score, Data Processing → Risk Traceability
ai-playbook 8.0 → 9.0 Remediation Roadmap, AI Ethics Score Projection
analyze 7.6 → 9.0 Orange Book fix, Weighted Health Score, Cross-Artifact Consistency
backlog 8.0 → 9.0 INVEST Story Quality Score, Definition of Ready

Standard pattern (36 commands) — Quality score + traceability matrix applied to the remaining 36 artifact-generating commands.

Consistency fixes

  • NFR-S-xxxNFR-SEC-xxx for security contexts in 6 commands
  • Added missing score instructions to 4 commands
  • Added missing traceability to 2 commands
  • Result: 48/48 commands have quality-checklist reference, quantitative score, AND traceability matrix

Rebase resolution (2026-04-19)

Branch was 27 commits behind main. Resolution summary:

Test plan

  • python scripts/converter.py → 408 files across 6 formats
  • Orange Book 2023 six-options present in risk command across all 6 output formats
  • NFR-SEC-xxx (not NFR-S-xxx) used for security
  • keep-coding-instructions preserved on plugin requirements/sobc, stripped from non-Claude formats
  • npx markdownlint-cli2 on tracked files: 0 errors
  • All 7 version files + plugin.json + gemini-extension.json at 4.7.0
  • CI lint green on rebased head
  • Spot-check /arckit:requirements, /arckit:risk, /arckit:sobc, /arckit:secure in a test repo before merge
  • After merge: tag v4.7.0, run ./scripts/push-extensions.sh, release workflow auto-creates GitHub Release

This was referenced Mar 27, 2026
This was referenced Mar 27, 2026
@tractorjuice tractorjuice force-pushed the feat/autoresearch-all-commands branch from 0d8f724 to c00dec4 Compare April 6, 2026 11:48
Rebases #265 onto main (v4.6.12), regenerates all 6 extension formats
from the optimised plugin commands, and bumps to v4.7.0.

## Content uplift (48 commands)

- **Orange Book 2023 alignment** (risk, analyze): replaced retired 4Ts
  with six treatment options (Avoid / Take-Increase / Retain / Change
  Likelihood / Change Consequences / Share), added Three Lines Model,
  tightened score boundaries
- **Green Book 2026 alignment** (sobc): Theory of Change, AST, VfM,
  monetisation split
- **NCSC CAF maturity** (secure): Maturity Summary, Remediation Roadmap,
  Principles-to-CAF traceability
- **MoSCoW + NFR sub-prefixes** (requirements): NFR-P / NFR-SEC /
  NFR-ACC / NFR-REL standardisation, Requirements Coverage Matrix
- **Quality score + traceability matrix** rolled out to all 48
  artifact-generating commands — every generated artifact now
  self-evaluates and includes cross-artifact consistency checks

## Conflict resolution

- requirements.md, sobc.md: merged PR body with #301 frontmatter
  (keep-coding-instructions: true preserved alongside effort: max)
- VERSION, plugin.json: accept HEAD (4.6.12 → 4.7.0 via bump-version.sh)
- doc-types.mjs: union merge (GLOS, MMOD codes)
- 4 extension trees (codex/opencode/gemini/copilot/paperclip):
  regenerated via scripts/converter.py

## Verification

- python scripts/converter.py → 408 files generated across 6 formats
- markdownlint-cli2 on all tracked .md → 0 errors
- All 7 VERSION files + plugin.json + gemini-extension.json at 4.7.0
- Orange Book 2023 content present in all 6 output formats
- keep-coding-instructions correctly stripped from non-Claude outputs

Closes #265

Co-Authored-By: Claude Opus 4.7 (1M context) <[email protected]>
@tractorjuice tractorjuice force-pushed the feat/autoresearch-all-commands branch from c00dec4 to a56b0e2 Compare April 19, 2026 05:28
@tractorjuice tractorjuice self-assigned this Apr 21, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant