fmk-plugins

C4-model documentation tracker with SHA propagation and staleness detection

On-demand codebase research. Parallel sub-agents produce SHA-tracked, line-referenced docs at docs/research/. Detects existing research on the same topic and updates instead of duplicating. Companion to docs.

Analyze and address pull request review comments. Fetches unresolved threads via GitHub GraphQL, evaluates each comment against real code, makes changes or drafts responses.

Research-backed implementation planning with a mandatory clarification dialog. Every open question (undefined terms, ambiguities, scope edges) must be resolved by user Q&A or codebase research before any plan body is written. Produces phased plans at docs/plans/ with SHA-tracked references and an embedded resolved-questions audit table.

Proof-based bug investigation. Forces explicit hypotheses, evidence-based verification, 5 Whys root cause analysis, and falsification check, with a written investigation log at docs/investigations/ before any code change is permitted.

Install an opinionated Claude Code statusLine into ~/.claude/settings.json. Layout: <folder> <model> effort: <level> ctx: <human-tokens> / <pct>%. Conditional segments, 1M / 200K context window fallback by model name, non-destructive merge, JSON validation, and dry-run smoke test.

Test-driven feature-build workflow for an unfamiliar codebase: map it into a navigable wiki and a C4 architecture model, cut the task into vertical walking-skeleton slices (each E2E-testable), drive every slice through /red → /green → /refactor with outside-in TDD, shift-left gating, and atomic feature-scoped commits, then /presubmit (full real gate + parallel multi-agent grading panel). Stack-agnostic.

Generate a complete, playable, visually beautiful game in one orchestrated multi-agent workflow. The orchestrator writes an immutable contract + an art-direction style bible, fans the build out across parallel implementers (with dedicated world/structure/character/fx/lighting art agents), runs a review → adversarial-verify → fix gauntlet, then actually runs the game, captures screenshots, and iterates against an art-director judge until it looks great. Stack-agnostic; browser/canvas games are the sweet spot. Composes the contract-first method with front-loaded art direction and a closed screenshot-judge feedback loop.

Turn a one-line product idea into a complete, running full-stack system (front + back) in one orchestrated multi-agent build. The architect expands the brief with taste, freezes a compiler-enforced contract (types + config + design system) after an adversarial pre-freeze gauntlet (≥3 refute-reviewers), product-directs every screen and endpoint in prose, fans out parallel disjoint-ownership agents (decomposition verified by check_plan.py), hardens via compile loop → multi-lens review → adversarial verification → mechanical gate, then closes the loop with Playwright-captured, scored product-design + UX judges (including a driven first-time-user task) iterating until every axis clears the bar. Generalization of the contract-first game method to front/back systems; model-tiered.

Open, inspect, and faithfully implement Figma .fig files locally — no Figma app, account, login, or network. Decodes the proprietary fig-kiwi binary (deflate schema + Zstandard data block, which breaks most third-party tools) to JSON; lists/queries screens, text, colors, design tokens; and extracts a complete z-ordered, scale-aware render spec of a frame (fills, image crop modes, gradients, blur/shadow effects, vector paths → SVG) so a design can be read or built in code pixel-accurately instead of approximated. Kiwi decoder is vendored, so it runs offline with no npm install.