jcdendrite
@jcdendrite
Public Skills
ready-for-review
by jcdendrite
Pre-handoff gate: runs verification, code review, syncs or opens a PR. TRIGGER when: handing off to a human reviewer — wrapping up a branch, "ship it" intent, before a multi-persona review (CISO + staff-* engineers) or /ultrareview, or on any push to a branch with an open PR, mid-iteration included. DO NOT TRIGGER when: no push or gh pr ready is attempted, or on the default branch.
claude-hook-review
by jcdendrite
Review and audit of Claude Code hooks (.claude/hooks/.sh and settings.json hook entries). TRIGGER when: reviewing a .claude/hooks/.sh script or settings.json hook entry, or designing a new hook. DO NOT TRIGGER when: editing permissions.allow rules (use review-permissions) or non-Claude hook systems.
pr-description
by jcdendrite
Author or sync the current branch's PR description. No open PR — draft the body to the standard here, check it, and report a file path for the caller to create the PR from. Open PR — fetch the body and verify it against branch state. Both modes read the body end to end for self-consistency, then flag per-commit narratives, stale content claims, TBD markers, and files missing from the body, and preserve coordination steps. Dispatched from /ready-for-review step 5 and the /handoff pre-write checklist; also run standalone.
review-loop-cost-audit
by jcdendrite
Decompose why a branch's review loop cost what it did and ran as many rounds as it did, by where the dollars concentrated and what triggered each expensive round, via a corpus-wide sweep or a single-branch deep audit. Reports shares, causes, and cheaper levers; the post-freeze flags (stuck loop, plan-grinding, gate-denial churn) are among several. For a narrative timeline of session prompts use transcript-narrative; for raw toolkit metrics use transcript-analysis.
transcript-analysis
by jcdendrite
Analyze Claude Code transcripts for cost, behavior, and review-activity statistics; the cost subcommand covers token cost, cache efficiency, and branch- or repo-scoped dollar cost. TRIGGER when: needing any statistic derived from Claude Code transcripts — cost, cache efficiency, model mix, subagent-vs-main split, test-failure convergence, session or branch history, a denial/friction census, or a per-session record of typed prompts and corrections. DO NOT TRIGGER when: the ask is a case study or annotated timeline built on top of these metrics (use transcript-narrative), or the number was already produced by this toolkit earlier in the session.
plan-it
by jcdendrite
Produces an implementation plan and hands off to /plan-review. TRIGGER when: asked for a plan or implementation strategy for work spanning multiple files or domains. DO NOT TRIGGER when: single-file tweaks, "just implement it" requests, or when a plan already exists (use /plan-review instead).
code-review
by jcdendrite
"Principal-engineer review before presenting code. Also the gate on git commit. TRIGGER when: code is about to be presented, a commit is pending, or the user asks for a review. skill-review, agent-review, and ai-instruction-and-memory-files run from inside this review, never instead of it; plan-review gates separately and also fires. DO NOT TRIGGER when: a fresh code-review-markers/ entry already covers the staged diff."
transcript-narrative
by jcdendrite
Produce a narrative case study / annotated timeline from Claude Code session transcripts — verbatim prompts, phase buckets, quantitative metrics, extracted lessons. For raw quantitative metrics use transcript-analysis.
plan-review
by jcdendrite
Review implementation plans before presenting to the user. A plan file in .claude/plans/ gates Write/Edit/MultiEdit/ExitPlanMode until this runs; triviality and user waivers do not release it. TRIGGER when: a plan is written or updated in .claude/plans/, or is about to be presented. DO NOT TRIGGER when: the plan lives outside .claude/plans/ (chat-level or /tmp drafts).
test-conventions-claude-config
by jcdendrite
Project-specific layer for /test-conventions, loaded only when writing or reviewing tests in the claude-config repo itself.
skill-review
by jcdendrite
Behavioral-equivalence audit of SKILL.md files (frontmatter, voice, length, duplication). Dispatched by name from /code-review and the require-skill-review hook — not auto-triggered by description.
handoff
by jcdendrite
Write a cross-session handoff file at ~/.claude/handoffs/<descriptive-slug>-handoff.md capturing goal, status, task list, next step, files modified, active markers, open questions, and resume command.
code-review-claude-config
by jcdendrite
Project-specific layer for /code-review, loaded only when reviewing changes in the claude-config repo itself.
config-environments
by jcdendrite
Multi-environment config design, credential isolation, and secrets provisioning. TRIGGER when: designing or reviewing env-var schemes, branching-on-environment credential selection, or secret-isolation strategies. DO NOT TRIGGER when: a process legitimately holds both env values simultaneously (migrations, test harnesses) or single-environment apps.
feature-flags
by jcdendrite
Which layer a runtime toggle belongs at, and whether a dedicated flag-management platform is warranted. TRIGGER when: deciding where a toggle lives (a deploy-time IaC variable, a field in a datastore the team already runs, or a flag platform), evaluating or adopting a feature-flag vendor, or reviewing a toggle whose lifespan and implementation weight are mismatched. DO NOT TRIGGER when: naming or sourcing a config value across environments (that is config-environments), auditing an already-implemented flag's test coverage or default-off semantics (code-review owns that), or any non-toggle env-var question.
subagent-delegation
by jcdendrite
Dispatch to a subagent vs inline. TRIGGER when: full check suite or full-project verification; broad codebase search; first exploratory read; 2nd/3rd Bash toward same question; delegating implementation; delegating a multi-step gate/review loop. DO NOT TRIGGER when: single-artifact targeted lookup (one file or value, not a multi-site sweep); comprehension read feeding your own writing/review/design; Edit/Write sequences where scope or content is still forming; the specific failure output or diff you reason over line by line.
ai-instruction-and-memory-files
by jcdendrite
Audit of CLAUDE.md, AGENTS.md, .claude/rules/.md, and Claude Code auto-memory files. TRIGGER when: authoring/reviewing a CLAUDE.md, AGENTS.md, .claude/rules/.md, or ~/.claude/projects//memory/ change, or deciding which surface a rule belongs in. DO NOT TRIGGER when: editing .lovable/.md, .agents/skills/*.md, .cursorrules, .github/copilot-instructions.md, README.md, or writing code.
brief
by jcdendrite
Produce a cold-start task briefing at ~/.claude/briefs/<slug>-task.md for a fresh session to pick up known, well-scoped work (abandoned PR, surfaced follow-up, ticket whose scope is settled). Distinct from /handoff, which captures mid-flight session state.
measure-check-output
by jcdendrite
Two-half procedure for grounding a harness-truncation or check-suite-size claim in a measured, dated figure instead of an assumed one. Half 1 re-probes the Claude Code Bash-tool output-truncation threshold and preview size by bisection. Half 2 measures a project's own documented check commands against that threshold, one project at a time, using two single-purpose Bash calls per command so no raw suite output enters context. Invoke by name when a harness-limit figure looks stale, or when a project needs its own check-suite-size baseline before writing an inline-vs-delegate rule that depends on one.
pr-description-claude-config
by jcdendrite
Project-specific layer for /pr-description, loaded only when authoring or syncing PR bodies in the claude-config repo itself.
git-state-safety
by jcdendrite
Ref inspection while mid-merge/rebase/cherry-pick without corrupting the index. TRIGGER when: examining another ref while the tree has merge/rebase/cherry-pick state or unresolved conflicts, or recovering from a bad merge already committed. DO NOT TRIGGER when: working on a clean tree or for force-push safety questions (use git-feature-branch-sync instead).
review-permissions
by jcdendrite
Security review of permissions.allow rules, bare permissions.deny entries, and permissions.defaultMode changes in settings.json. TRIGGER when: permissions.allow changes, a bare permissions.deny entry is added, or permissions.defaultMode changes, in any .claude/settings.json; or user asks to review permission/allow rules. DO NOT TRIGGER when: reviewing hook entries (use claude-hook-review); other settings.json fields (env, model, theme); Bash behavior unrelated to permissions scoping.
git-feature-branch-sync
by jcdendrite
Rebase vs merge a feature branch; force-push safely. TRIGGER when: syncing a feature branch with the default branch or deciding whether and how to force-push. DO NOT TRIGGER when: routine work on a clean branch, operating on shared/default branches (never force-push them), or mid-merge/rebase/cherry-pick state (use git-state-safety instead).
sql-query-conventions
by jcdendrite
Read-path SQL conventions for raw SQL, PostgREST, and SQL-shaped ORMs. TRIGGER when: writing or modifying a SELECT query, PostgREST .from()/.select() chain, list-returning ORM call, or query-shape / pagination design. DO NOT TRIGGER when: writing write-path (INSERT/UPDATE/DELETE), DDL-only migrations, test fixtures, or document/KV stores.