dotfiles

Address pending PR review feedback through verified triage, small fix phases, implementation, validation, and evidence-backed replies. Manual invocation only.

Autonomous iteration loop for a measurable goal: review, ideate, modify, verify, keep or rollback, repeat until interrupted or capped.

Monitor a PR's CircleCI pipeline, diagnose failures, apply scoped fixes, push when requested, and continue until green or blocked.

Stage and commit changes in logical groups using the project's git message style, without mixing unrelated user changes.

Create or update a GitHub PR with concise description, review guidance, triggered specialty reviews, and focus areas.

Build a daily work brief from yesterday/off-hours activity plus today's Linear, Calendar, Gmail, and Notion context; update Notion and produce standup/checklist.

Consolidate today's Notion entry, summarize Linear/work activity, and rewrite actions, decisions, and notes into a clean end-of-day record.

Capture a discovered failure pattern or correction as a gotcha for an existing skill.

Discover and recommend available personal skills for a requested workflow, grouped by purpose with usage guidance.

Convert the current conversation into an implementation path, choosing quick or full planning based on scope and risk.

Log an off-hours on-call incident to the daily Notion doc with a timeline useful for follow-up or postmortem.

Append concise session accomplishments to today's Notion work log.

Compare research, specs, plans, and related artifacts to find contradictions, coverage gaps, scope drift, and implementation-readiness risks.

Architecture review of a PR, document, or code area for boundaries, coupling, cohesion, dependency direction, and maintainability.

Surface ambiguities, contradictions, hidden assumptions, and underspecified areas in specs or research before planning.

Design evaluation plans for AI/LLM features: datasets, scorers, baselines, success criteria, and regression strategy.

Execute an approved plan phase by phase. Dispatches one implementation-executor at a time, enforces RED -> GREEN -> VALIDATE, re-verifies results, and stops on repeated failure.

Investigate production, runtime, CI, or test issues by gathering evidence from logs, metrics, traces, code, and CI, then ranking hypotheses.

Synthesize current session state into a prioritized action plan after research, validation, review, or divergent discussion.

Design observability and monitoring for planned code changes: metrics, traces, spans, logs, dashboards, and actionable alerts.

Create a detailed implementation plan with mechanically verifiable success criteria, phase boundaries, tests-first steps, and explicit non-goals.

One-pass implementation workflow for small, well-understood changes with lightweight research, focused TDD, validation, and self-review.

Deep codebase research with verified findings. Uses focused discovery agents, cross-checks claims against code, saves a durable research artifact, and records assumptions.

Rigorous code review for local diffs or GitHub PRs. Uses a high bar for REQUEST_CHANGES: only Critical merge-blocking issues should block a PR; approve only when requirements are satisfied.

Refine vague ideas, bugs, or rough requests into scoped technical-product specs with problem statement, boundaries, acceptance criteria, and open decisions.

Execute a manual E2E test plan in the browser, record evidence, and format results for the user or PR.

Produce a manual E2E test plan from a ticket, PR, or code change with scenarios, setup, expected outcomes, and risk coverage.

Validate work against a plan or current session. Runs mechanical checks, verifies claims against code, repairs local failures when safe, and reports residual risk.

Run one checkpointed delivery stage at a time: research, spec, clarify, plan, observe, analyze, then the gated atomic implement -> validate -> review block. Never jump straight to implementation unless the workflow ledger explicitly marks all prior stages complete.

Deep performance review: query plans, index coverage, load impact, caching strategy, resource use, and scaling risks.

Analyze Linear milestone issues, sequence dependencies, identify critical path, and optionally update issue relationships.

Publish prepared PR review to GitHub inline comments, thread replies, review body via gh api.

Deep audit of test quality, coverage fidelity, flakiness risk, assertions, mocks, and whether tests actually catch intended bugs.

Execute a quick-plan one phase at a time using quick-implement-agent for TDD or direct-edit phases.

Create a small executable quick-plan for simple changes, with TDD or direct-edit phases and mechanical success criteria.

Audit a PR, ticket, spec, or local change against stated requirements and acceptance criteria.

Deep security audit for code changes or areas: auth/authz, data exposure, injection, secrets, dependency risk, and OWASP-style issues.

Plan parallel milestone work for multiple developers by analyzing issue surfaces, dependencies, sequencing, and merge-conflict risk.

Filter previous findings through an importance bar so only issues worth raising or fixing survive.

Update outdated dependencies across supported package managers, handling safe updates first and flagging breaking changes.