flurdy/agent-skills

flurdy/agent-skills

Skill Claude Code

Sort, dedupe, and audit Claude settings.json / settings.local.json at user/project level. Flags risky permissions, broken references, glob-subsumed entries, and cross-file duplicates. Mechanical fixes auto-apply; judgment calls become triage items.

flurdy/agent-skills

Skill Claude Code

Read-only activity recap for today or the previous workday across commits, PRs, Jira touches, and Beads in validated repositories; same-day mode adds current-session context.

flurdy/agent-skills

Skill Claude CodeCodex

Read-only current-session and UTC-week token telemetry dashboard for Pi, Claude Code, Codex, and optional OpenRouter management analytics.

flurdy/agent-skills

Skill Claude Code

Portable pre-PR quality gauntlet: cleanup, verification, craft, correctness, security, and optional independent reviews. Binds every gate to the final scope, reports missing coverage, and caps fix/review passes at two.

flurdy/agent-skills

Skill Claude Code

Portfolio-wide drift sweep across Jira, beads, and GitHub PRs — flags status drift, orphan work, parent-moved beads, and stale items. Read-only; use for a "where is everything" reconciliation, unlike /landscape's passive snapshot.

flurdy/agent-skills

Skill Claude CodeCodex needs its repo

Integrate Trello boards with Beads — pull triage cards into beads, sync closed beads back to Trello. Use for project management bridging between Trello and Beads.

flurdy/agent-skills

Skill Claude Code

Create or refine beads from raw requests, Jira, or explicit human-decision handoffs. Investigates relevance and duplicates; delegates approved implementation plans to plan-to-backlog.

flurdy/agent-skills

Skill Claude Code

Verify explicit requirements and coverage against a fixed implementation scope using discovered repository-native gates. Reports missing, failed, or stale evidence; never fixes code or changes tracking.

flurdy/agent-skills

Skill Claude Code

After a push or merge, watch a CircleCI + FluxCD deploy until it lands — CI green, then the k8s Deployment image tag moves and pods go ready — then run a read-only smoke test scoped to the change. Goal-terminating loop; stops on landing or failure.

flurdy/agent-skills

Skill Claude Code

Watch the user's open GitHub PRs for normalized feedback, validate each new or edited actionable item once, and render a bounded decision queue. Read-only by default; attended mode pauses only when the queue needs acknowledgment.

flurdy/agent-skills

Skill Claude Code

Start a recurring PR status dashboard — runs /pr-status on an adaptive cadence (fast when CI is in flight, backing off when settled) until end of day. Unattended: renders tables and suggested next actions, never prompts or blocks.

flurdy/agent-skills

Skill Claude Code

Start a recurring release-gatekeeper loop — runs /release-manager on an adaptive cadence (fast when something is in flight, backing off when settled) until end of day. Prompts to push / defer / cancel as services become ready; runs attended.

flurdy/agent-skills

Skill Claude Code

Watch for direct GitHub review requests, run one bounded repository-qualified review at a time, and pause for private, draft-only, defer, or separately confirmed external dispositions.

flurdy/agent-skills

Skill Claude Code

Choose and launch the appropriate rollout watcher. Prompts between implemented deployment stacks unless an explicit selector is supplied, then delegates without weakening stack-specific safety.

flurdy/agent-skills

Skill Claude Code

End-of-session handoff — summarise today's commits, PRs, and beads, warn about uncommitted/unpushed work (across all repos in a multi-repo workspace, and in worktrees), and emit a paste-ready resume block. Run before leaving the client.

flurdy/agent-skills

Skill Claude Code

Read-only previous-workday recap. Alias for /today --previous-workday; the shared workflow selects Friday when run on Monday.

flurdy/agent-skills

Skill Claude Code

Load before authorized code changes: features, fixes, refactors, tests, and mechanical edits, including discussion-to-coding transitions. Read separately from edits. Skip read-only requests and implementation already owned by a specialist.

flurdy/agent-skills

Skill Claude Code

Rebase the current branch onto an updated main, an updated stacked parent, or main after the parent PR merged. Proves the child-only commit range before rewriting anything; explicit gates for dirty trees, tests, force-push, and PR retargeting.

flurdy/agent-skills

Skill Claude Code needs its repo

Separately confirmed manifest reconciliation, scoped config synchronization, workload restart, or evidence-backed rollout acknowledgement. Requires explicit invocation; never runs inside a release watch tick.

flurdy/agent-skills

Skill Claude Code

After a merge, watch the GitHub Actions deploy until the gating job lands, then run a scoped smoke test against staging (browser for UI, GET for read-only API). Goal-terminating loop — stops when the deploy lands and the smoke completes, or fails.