Commit graph

141 commits

Author SHA1 Message Date
dan 64c1c9350e bd sync: 2026-01-01 17:36:29 2026-01-01 17:36:29 -08:00
dan fb882a9434 feat: add ops-review skill with Phase 1 lenses
Multi-lens review skill for operational infrastructure (Nix, shell,
Docker, CI/CD). Modeled on code-review with linter-first hybrid
architecture.

Phase 1 lenses (core safety):
- secrets: credential exposure, Nix store, Docker layers, CI masking
- shell-safety: shellcheck-backed, temp files, guard snippets
- blast-radius: targeting/scoping, dry-run, rollback
- privilege: least-privilege, containers, systemd sandboxing

Design reviewed via orch consensus (sonar, flash-or, gemini, gpt).
Lenses deploy to ~/.config/lenses/ops/ via home-manager.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2026-01-01 17:36:24 -08:00
dan 503053638a docs: update doc-review skill to match implementation
- Remove Vale references (tool has own deterministic rules)
- Document all 12 deterministic + 7 LLM checks
- Add missing CLI flags (--llm, --format, --model)
- Add concrete example session with output

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-31 22:55:38 -05:00
dan b8f7928db5 refactor: single source of truth for lenses
- Remove top-level lenses/ dir (was stale with only 4 lenses)
- Update ai-skills module to source from skills/code-review/lenses/
- Fix visit.py: extract DEFAULT_WAIT_MS, use specific exception handlers

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-31 22:34:28 -05:00
dan d4a0bdb158 feat: add 5 new code-review lenses
New lenses: security, error-handling, coupling, boundaries, evolvability
Updated SKILL.md and lenses/README.md to reflect 9 total lenses

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-31 18:18:01 -05:00
dan 34afa86b77 docs: worklog for vision benchmark, orch patterns, claude-search 2025-12-29 20:45:01 -05:00
dan dd45e206ef bd sync: 2025-12-29 20:41:15 2025-12-29 20:41:15 -05:00
dan 7e3ba83101 feat: add claude-search prototype for conversation history
- Indexes all Claude Code sessions (summaries + metadata)
- Searches by keyword in summaries and first message
- Shows date, project, message count, uuid
- Usage: claude-search <query> or --rebuild
2025-12-29 18:35:50 -05:00
dan ef28102acd docs: add conversational patterns to orch skill
- Session-based multi-turn conversations
- Session inspection commands
- Cross-model dialogue pattern
- Iterative refinement pattern
- When to use conversations vs consensus table
- Combined explore-then-validate pattern
2025-12-29 15:34:09 -05:00
dan bd83887669 research: vision model UI understanding benchmark
Tested Claude Opus 4.5 on btop and GitHub screenshots.
Findings: excellent text/state/layout, approximate coordinates.
Recommendation: hybrid AT-SPI + vision approach.
2025-12-29 15:26:13 -05:00
dan be6457e3b4 bd sync: 2025-12-29 15:05:07 2025-12-29 15:05:07 -05:00
dan 93d582576b bd sync: 2025-12-29 14:37:43 2025-12-29 14:37:43 -05:00
dan c886c529bf bd sync: 2025-12-29 13:55:41 2025-12-29 13:55:41 -05:00
dan 324b4a6fa3 docs: worklog for issue triage and playwright-visit implementation
Session covered:
- 9 issues closed (design decisions + implementation)
- playwright-visit skill created and tested
- READMEs added for web-search and web-research
- ADR-001 parked, multiple design questions resolved

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-29 01:22:19 -05:00
dan ad15afecf0 fix(playwright-visit): use domcontentloaded instead of networkidle
networkidle times out on heavy sites like nytimes.com.
domcontentloaded + wait parameter is more reliable.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-29 01:15:34 -05:00
dan 076588a34d bd sync: 2025-12-29 00:09:55 2025-12-29 00:09:55 -05:00
dan d24bedcab3 feat: add playwright-visit skill for browser automation
- visit.py CLI with subcommands: screenshot, text, html, pdf
- Uses system chromium on NixOS (no browser download)
- Fresh profile each run (no cookies/history)
- flake.nix provides playwright devShell
- Options: --wait, --full-page

Useful for JS-heavy sites where WebFetch fails.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-29 00:09:50 -05:00
dan efb7cdaffc bd sync: 2025-12-28 23:27:55 2025-12-28 23:27:55 -05:00
dan c7c6bbf796 docs: park ADR-001 skills-molecules integration
Current simpler approach working well:
- Skills as standalone entrypoints
- Agent judgment sufficient for invocation
- Molecules not actively used

Revisit when complex orchestration is needed.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-28 23:27:41 -05:00
dan 714e2f208a bd sync: 2025-12-28 22:46:28 2025-12-28 22:46:28 -05:00
dan 9f9f563d1e bd sync: 2025-12-28 22:37:56 2025-12-28 22:37:56 -05:00
dan 6ad2f0a97c docs: add README.md for web-search and web-research skills
Human-facing documentation with:
- Prerequisites and installation
- Usage examples
- Backend options (web-research)
- Cross-references between related skills

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-28 22:37:47 -05:00
dan 7477482cca bd sync: 2025-12-28 22:28:50 2025-12-28 22:28:50 -05:00
dan 6a9b2bef50 bd sync: 2025-12-28 22:25:38 2025-12-28 22:25:38 -05:00
dan ac22f23d5b bd sync: 2025-12-28 22:22:57 2025-12-28 22:22:58 -05:00
dan e012b168b5 bd sync: 2025-12-28 22:21:12 2025-12-28 22:21:12 -05:00
dan f3d7c5bc2d docs: resolve web search design questions
Context changed - Claude Code now has built-in WebSearch/WebFetch.

Decisions:
- Keep two separate skills (web-search, web-research)
- web-search: OpenCode only (Claude Code has native search)
- web-research: Both (deep research beyond built-in capability)
- Source filtering via WebSearch allowed_domains, not skill feature
- Summaries by default, not raw links

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-28 22:21:05 -05:00
dan e209db9230 docs: worklog for niri fixes, opencode research, readme update
Session covered:
- skills-m21: niri-window-capture robustness improvements
- skills-czz: OpenCode agents research
- skills-4yn: screenshot-latest deployment
- skills-a23: README update with all 14 skills

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-28 22:10:24 -05:00
dan eff02f2def bd sync: 2025-12-28 22:08:08 2025-12-28 22:08:08 -05:00
dan cf5033f665 docs: update README with all 14 skills
Replace verbose 2-skill list with table of all skills.
Shows deployed (5), available (8), and development (1) skills.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-28 22:08:01 -05:00
dan 23430b89f7 bd sync: 2025-12-28 20:55:18 2025-12-28 20:55:18 -05:00
dan ad49fd18d2 docs: mark screenshot-latest and niri-window-capture as deployed
- screenshot-latest: decided to deploy (low risk, useful)
- niri-window-capture: already deployed, code review fixes applied

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-28 20:55:10 -05:00
dan a0a386187e bd sync: 2025-12-28 20:49:03 2025-12-28 20:49:03 -05:00
dan 3ca79bb3bd docs: document OpenCode agents research
- Built-in agents: Build (full access) and Plan (read-only)
- Custom agents support prompts, model overrides, tool control
- Skills are global, agents access via skill() tool
- Permission control per-agent via opencode.json
- OpenCode supports Claude-compatible skill paths

Closes skills-czz

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-28 20:48:58 -05:00
dan b1cb619f7a bd sync: 2025-12-28 20:16:54 2025-12-28 20:16:54 -05:00
dan 62e5d7f7dc fix(niri-window-capture): add robustness improvements
- Add dependency checks for niri, jq, logger at script startup
- Validate and create screenshot directory if missing
- Improve error handling for niri command failures
- Add audit logging to capture-all-windows.sh
- Use variable for screenshot directory path
- Cache window list query in capture-all-windows.sh

Addresses recommendations from code review (skills-m21)

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-28 20:16:47 -05:00
dan e603e09862 feat: add code-review to availableSkills 2025-12-28 20:16:38 -05:00
dan 8f616bbb03 bd sync: 2025-12-28 00:07:09 2025-12-28 00:07:09 -05:00
dan fb5e3af8e1 docs: worklog for code-review skill creation and worklog cleanup 2025-12-28 00:06:38 -05:00
dan 4b72e6fc2e refactor(worklog): consolidate git commands into script
- Move all git context gathering into extract-metrics.sh
- Script now outputs: branch, status, commits, files, lines
- Remove redundant Helper Scripts entry
- Semantic compression refs already removed in previous commit

98 -> 88 lines (-10%)
2025-12-27 10:11:56 -05:00
dan 862494d515 bd sync: 2025-12-27 10:06:41 2025-12-27 10:06:41 -05:00
dan 82445dd839 refactor(worklog): consolidate skill prompt
- Remove inline section list, reference template directly
- Merge Guidelines and Remember into single Principles section
- Extract skill directory path, use relative references

127 -> 98 lines (-23%)
2025-12-27 10:06:02 -05:00
dan 3edc3f6502 docs: add code-review to skills list 2025-12-26 23:48:37 -05:00
dan aec0704e69 bd sync: 2025-12-26 23:22:52 2025-12-26 23:22:52 -05:00
dan 5ebae8f86e feat: add /code-review skill with bundled lenses 2025-12-26 23:22:51 -05:00
dan 2103e0994d docs: worklog for multi-lens code review workflow testing 2025-12-26 02:04:08 -05:00
dan 1e645151e6 bd sync: 2025-12-26 01:57:15 2025-12-26 01:57:15 -05:00
dan 9624873dc2 update: code-review proto with LLM-in-the-loop pattern
- Emphasize agent judgment at filing step (not mechanical parsing)
- Add exploration step for directory targets
- Document synthesis across lenses
- Clarify prioritization and grouping guidance

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-26 01:51:10 -05:00
dan fb15000877 refactor: restructure for cross-repo deployment
- Rename reviews/ to lenses/ (clearer terminology)
- Add workflows/ for beads proto templates
- Extract code-review proto to workflows/molecules.jsonl
- Update ai-skills.nix module:
  - Add enableLenses option (deploys to ~/.config/lenses/)
  - Add enableWorkflows option (deploys to ~/.beads/molecules.jsonl)
  - Derive repoRoot from skillsPath for sibling directories
- Update lens paths in proto to use deployed location

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-26 01:18:19 -05:00
dan 6878783bf1 feat: add review lenses for multi-perspective code analysis
Four focused review prompts for use with orch multi-model consensus:
- bloat.md: file size, function length, complexity
- smells.md: code smells, readability, naming
- dead-code.md: zombie code, unreachable, obsolete
- redundancy.md: duplication, parallel systems, YAGNI

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-25 01:52:35 -05:00