tansuasici
UserDrop-in starter templates that make Claude Code behave like a disciplined staff engineer instead of an eager intern.
Categories
Indexed Skills (61)
accessibility-audit
Audit UI code for WCAG 2.1 AA compliance — semantics, keyboard navigation, color contrast, and ARIA. Use when auditing accessibility or the user mentions a11y, screen readers, WCAG, or keyboard support. For visual/design quality use /design-review instead.
architecture-review
Review codebase architecture for SOLID violations, dependency health, module boundaries, and structural issues. Use when assessing whole-system structure or coupling. For single-module depth use /deepening-review; for a concrete refactor plan use /refactoring-guide.
code-quality-audit
Audit code for smells, error-handling gaps, and maintainability issues with actionable fixes. Use for a general code-quality pass not tied to project rules. To audit against the project's golden-principles.yaml use /quality-audit instead.
constitution
Author or extend the project's golden-principles.yaml — interactively for new repos, or by inferring and confirming candidates from existing code. Use when defining project coding principles or quality gates. Pairs with /quality-audit, which reads the file.
dead-code-audit
Detect unused code — unreferenced functions, dead imports, orphan files, and unreachable branches. Use when hunting dead code before a cleanup or release, or when the user suspects code is unused.
debug
Systematic root-cause debugging with evidence-before-fix enforcement and regression-test generation. Use when chasing a bug, failing test, error, crash, or unexpected behavior — before attempting a fix.
deepening-review
Surface shallow, pass-through modules as deepening candidates, then grill the chosen one interactively (Ousterhout depth/seam paradigm). Use when modules feel thin, fragmented, or hard to navigate. For whole-system structure use /architecture-review.
dependency-audit
Audit dependencies for security vulnerabilities, outdated versions, license risks, and bloat. Use when reviewing third-party packages, before a release, or when the user asks about CVEs, updates, or supply-chain risk.
design-review
Review UI implementation for design consistency, visual quality, AI-slop detection, and responsive behavior. Use when evaluating how a built UI looks and feels. For accessibility use /accessibility-audit; to build new components use /ui-component-builder.
doc-gardening
Cross-check docs/ against current code — flag stale file paths, broken cross-references, removed APIs, and outdated examples. Use when docs may have drifted after code changes. For doc completeness and quality use /documentation-audit instead.
documentation-audit
Audit documentation quality — inline comments, API docs, README completeness, and doc-code sync. Use when assessing whether docs are good and complete. To catch docs that have gone stale versus the code use /doc-gardening instead.
feature-cycle
End-to-end orchestrator chaining spec → plan → implement → verify → review → ship from a local spec — the whole CLAUDE.md lifecycle as one command. Use to drive a complete feature from a ready spec. To validate what to build first use /office-hours.
harness-init
Scaffold a docs/ harness (ARCHITECTURE, DESIGN, PLANS, QUALITY_SCORE, RELIABILITY plus design-docs/, exec-plans/, references/). Idempotent — never overwrites. Use once when adopting the harness pattern in a repo lacking docs/ structure.
interface-design
Design It Twice — spawn 3-4 parallel sub-agents producing radically different interfaces for a module, then compare on depth, locality, and seam placement. Use before committing to a new interface or public API.
lesson-refresh
Periodic maintenance of tasks/lessons/ — decide keep/update/promote/encode/archive per lesson by relevance, recency, and prior encoding. Use to prune lesson rot. To recall dormant lessons for the current task use /lesson-resurface instead.
lesson-resurface
Surface archived or superseded lessons whose topic tags match the current task — returns pointers (paths only), never bodies. Use when a task touches an area past sessions may have covered, to recover dormant context without bloating boot.
office-hours
Pre-coding product validation using forcing functions — clarify what to build and why before any code. Use at the start of a feature when scope or value is fuzzy. To turn a validated idea into a spec use /shape-spec; to run the build use /feature-cycle.
performance-audit
Identify performance bottlenecks — rendering, startup, memory, and I/O — across any stack. Use when the app is slow, before perf-sensitive releases, or when the user mentions latency, jank, or memory growth.
project-health-report
Comprehensive whole-project health report across code quality, architecture, testing, dependencies, and documentation. Use for a breadth-first snapshot of the entire repo. For PR-scope diffs use /review-pipeline; for one dimension use the specific /*-audit.
pulse
Time-windowed pulse report on what shipped, broke, was learned, and is open — saved to tasks/pulses/ as a timeline. Use to summarize a recent period of work. Distinct from /retro (process) and /project-health-report (whole-project code health).
refactoring-guide
Fowler-based refactoring recommendations with specific techniques, risk assessment, and step-by-step plans. Use when code needs restructuring and you want a safe execution plan. To first locate structural issues use /architecture-review.
references-sync
Sync llms.txt-style dependency reference docs into docs/references/ so the agent has library docs in-context — scans package manifests and fetches known vendor references. Use when working with a third-party library whose docs aren't yet local.
review-pipeline
Run multiple audits in parallel over a PR-scope diff, dedupe across auditors, and produce a confidence-gated report. Use to review a change or PR before merge. Distinct from /project-health-report (whole-project, breadth-first).
shape-spec
Create a timestamped feature-spec folder with structured plan, decisions, and references for multi-session features. Use to capture a spec before building something non-trivial. To validate the idea first use /office-hours; to execute it use /feature-cycle.
ship
Full deployment pipeline — tests, coverage audit, CHANGELOG generation, bisectable commits, and PR creation. Use when a change is verified and ready to ship as a pull request.
skill-extractor
Extract non-obvious knowledge discovered during a session into a reusable SKILL.md. Use right after learning something hard-won — a framework quirk, a verified fix. To generate skills proactively from stack analysis use /skill-generator instead.
skill-generator
Generate project-specific coding skills by analyzing tech stack, architecture, and constraints. Use when bootstrapping skills for a new project. To capture a single insight discovered mid-session use /skill-extractor instead.
testing-audit
Audit the test suite for coverage gaps, test quality, flaky tests, and strategy alignment. Use when assessing test health before a release, or when tests feel weak, slow, or flaky.
ui-component-builder
Build production-ready UI components with accessibility, loading/empty/error states, responsive behavior, and reusable props — not retrofitted polish. Use when creating a new component. To review an existing UI use /design-review instead.
capabilities
Summarize what the kit makes available in this project — user-invocable skills, review agents, active hooks, and enabled modules — as a one-shot onboarding briefing read live from disk. Use when you start in a kit-enabled project or ask "what's available here / what can I do". Do NOT use for install health (that's scripts/doctor.sh) or terminal skill listing (npx @tansuasici/claude-code-kit skills).
mcp-audit
Audit the project's configured MCP servers against the trust allowlist (.claude/mcp-allowlist.txt) that .claude/hooks/mcp-gate.sh enforces. Lists each configured server, flags servers that are NOT allowlisted (so their tool calls would be blocked or, if the gate is off, run untrusted), and surfaces prompt-injection / supply-chain risk. Use when adding or reviewing MCP servers, before turning on the gate, or when an mcp__* tool call was blocked. Do NOT use to install MCP servers or to run their tools.
web-read
Extract clean, readable markdown from a web page with the Defuddle CLI — strips nav, ads, and boilerplate to cut tokens versus WebFetch. Use when given a URL to read or analyze (docs, articles, blogs, changelogs, RFCs). Do NOT use for .md or JSON/API URLs — use WebFetch directly.
wiki-briefing
Morning briefing from the knowledge wiki — recent activity, new sources, open actions, and key updates
wiki-ingest
Ingest a new source into the knowledge wiki — read, summarize, cross-reference, and update index
wiki-lint
Health-check the knowledge wiki — find contradictions, orphans, missing pages, and stale content
note
Append a timestamped mid-session note to the session journal — finding/decision/summary tags for across-compaction memory. Use when you discover something the session must remember past /compact. Folded to the handoff at session end.
retro
Weekly retrospective with session analytics, LOC metrics, pattern detection, and persistent history. Use to reflect on how work went over a week. For what shipped/broke as a timeline use /pulse; for raw session numbers use /scorecard.
scorecard
Aggregate recent session scorecards from reports/session-audit.log into per-session and windowed summaries — pure numbers, fed by the SessionEnd hook. Use to see quantitative session trends. For narrative reflection use /retro instead.
abstract
Draft or tighten the abstract as a contract — every number and claim in it must appear identically in the body, within the venue's word limit, calibrated to what the paper actually shows
citation-audit
Deep manual bibliography health check — dangling \cite keys, orphan .bib entries, duplicates, malformed/placeholder DOIs, missing required fields, inconsistent author/journal formatting, and \ref↔\label integrity
claim-check
Walk every substantive claim in a section, classify it (cited / author's-own / common-knowledge / UNSUPPORTED), verify the citation licenses the claim's verb and quantifier, and report supported/overstated/uncited counts
cover-letter
Draft a concise, honest editor cover letter from MANUSCRIPT_MAP — states the contribution, fit to the venue, and required declarations, with no fabricated significance or invented metrics
gap-finder
Breadth-first scan of a draft for unsupported and uncited claims and missing-evidence gaps — classify every claim, list what is UNCITED/UNSUPPORTED, and for true gaps emit search directions (keywords, venues, citation-chaining), never a fabricated source
journal-fit
Assess a manuscript's fit to a target journal/venue — scope, novelty bar, length & structure limits, reference style (ACS/IEEE/APA/Nature), audience, article types — and output a fit score with reasoning and a prioritized gap list
lit-briefing
A short "since last time" briefing from vault/log.md + vault/index.md — recently ingested sources, open follow-ups and citation-chaining leads, and gaps relative to MANUSCRIPT_MAP.md → Thesis (themes the argument needs that the library lacks). Read-only; outputs a tight briefing whose gaps become search directions for /literature-review. Never fabricates a source to fill a gap.
lit-ingest
Ingest ONE raw source (a sources/ PDF or .txt path, or pasted text) into the Literature Vault — read it, write a vault/summaries/<bibkey>.md annotated page with claims+locators and verbatim quotes, propose a references.bib entry extracted FROM the document (never fabricated), cross-reference concepts/entities, and update index + log. Never modifies sources/; flags uncertain metadata as [VALUE — verify].
lit-lint
Health-check the Literature Vault — find cross-source contradictions, orphan pages (a summary with no matching references.bib key, or a .bib key with no summary), broken [[wikilinks]], claims missing locators, metadata-only pages never upgraded, stale entries, and quotes without page numbers. Read-only/advisory — reports a categorized findings list with fixes; never edits sources or auto-rewrites the vault.
literature-review
Synthesize a related-work / literature-review section from the project's OWN library (references.bib + sources/ + vault) — thematic, gap-driven, calibrated, real citations only. Proposes concrete search directions for gaps without fabricating a single source.
manuscript-cycle
End-to-end orchestrator — runs the whole CLAUDE.md research lifecycle for a section or manuscript as one command (outline → ground → draft → verify → review → revise), halting on any gate failure.
methods-review
Reproducibility check of the Method(s) section against agent_docs/reproducibility.md — enumerate what an independent team needs to rerun the work, check the section, and flag every missing ingredient as a pass/GAP checklist
outline
Turn a thesis + target venue into a claim-driven IMRaD outline — each section is the one claim it establishes plus the evidence needed and a word budget — ready for MANUSCRIPT_MAP.md, flagging evidence missing from the library
peer-review
Run a full simulated peer review — dispatch the peer-reviewer and integrity-reviewer agents over the manuscript, dedupe their findings, and produce a referee report with a recommendation. Pre-empt Reviewer 2 before submission
plain-language-summary
Draft a lay / plain-language summary for a general audience, funder, or a journal's required PLS — accessible to a non-specialist without distorting the science, with a claim-by-claim fidelity check against the paper
reference-format
Convert or normalize the bibliography's citation style via the toolchain (biblatex/biber, CSL via pandoc, bibtex .bst) — deterministic, presentation-only, never altering the underlying facts and never inventing a missing field
response-to-reviewers
Draft a point-by-point response letter — quote each reviewer comment, state the change made and its exact location, stay courteous, and NEVER claim a change not actually made. Produces the letter plus an edit checklist, logging recurring critiques to tasks/reviews/
review-resurface
Surface dormant reviewer-feedback notes from tasks/reviews/ (and _archive/) whose applies_to topics match the current task, returning POINTERS ONLY — paths plus frontmatter, never bodies. Recovers a past reviewer rule relevant to the work at hand without bloating Tier-1 boot. Delegates the deterministic scan and scoring to scripts/review-resurface.sh.
stats-check
Run the agent_docs/statistics.md checklist over a Results section — flag bare p-values, missing N/effect-size/CI, causal overclaim on observational data, multiple-comparison issues, and numbers that disagree across text/tables/abstract
submission-pipeline
Pre-submission review battery — runs the peer-reviewer, integrity-reviewer, and fact-checker agents plus the deterministic audits in parallel over the whole manuscript, dedupes, confidence-gates, and saves one go/no-go report.
note
Append a timestamped tagged line (finding / decision / summary) to the session journal at .hook-state/session-journal.md, so a discovered fact, a settled choice, or a transient breadcrumb survives context compaction in a long writing session. Delegates to scripts/note.sh.
retro
Windowed retrospective on the writing process — what shipped (sections drafted, commits over the window), recurring reviewer critiques (tasks/reviews/ by applies_to frequency), what was learned (journal findings folded into handoffs), and what is still open (tasks/todo.md). Saves a timestamped retro to tasks/ as persistent history. Process, not telemetry.
scorecard
Aggregate reports/session-audit.log (one JSON line per session, written by the SessionEnd hook) into a per-session table plus a windowed summary — citation-gate pass rate, which guardrail hooks fired, SKIP_QUALITY_GATE bypasses, and session durations. Pure telemetry; flags trends, does not judge process.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.