Original, hand-written Claude Code skills — plain markdown, no dependencies, no telemetry. Two paid packs, one free tool.
Five skills + one subagent for the pre-ship judgment calls: migration safety, real changelogs, blameless postmortems, dependency-upgrade risk, flaky-test diagnosis.
$15
Buy →Five skills + one subagent for when something's already broken: root-cause tracing, perf-regression hunting, log triage, memory-leak detection, concurrency-bug finding.
$15
Buy →Turn git log into a categorized changelog with a suggested semver bump. No API key, no network call, no LLM.
Guide: how to generate a changelog from git log without an LLM →
Guide: diagnosing flaky tests — a checklist for the four real causes →
Guide: safe database migrations — a checklist for lock, rollback, and rollout risk →
Guide: dependency upgrade risk — why the changelog and the semver bump both lie →
Guide: blameless postmortems — a checklist for the three ways they quietly fail →
Guide: root-cause tracing — why the first plausible cause is usually wrong →
Guide: perf regression hunting — the five shapes that cover most real-world slowdowns →
Guide: error log triage — why volume is not impact →
Guide: memory leak hunting — the six mechanisms behind unbounded growth →
Guide: concurrency bug finding — name the interleaving, not "might race" →
Guide: writing a CLAUDE.md that actually changes behavior →
Guide: skill, subagent, or slash command? Picking the right container →
Guide: hook or skill? When automation needs to be guaranteed →
Guide: git worktrees for parallel AI coding agents →
Guide: running Claude Code headless in CI — permissions when no one can click "allow" →
Guide: reviewing an AI agent's diff — what a green CI run won't catch →
Guide: secrets and credentials in an AI agent's shell — where they actually leak →
Guide: designing tools for AI agents to call — idempotency, dry-run, and blast radius →
Guide: what to log when an agent acts autonomously — building a reconstructable audit trail →
Guide: designing reversible actions for autonomous agents — undo, not just retry-safety →
Guide: scoping an agent's permissions — why "can run any command" is the bug →
Guide: prompt injection defense — when the content an agent reads is also an instruction channel →
Guide: sandboxing code an agent writes and runs — isolation, not just permission scope →
Guide: trust boundaries between agents — when a subagent's output becomes another agent's input →
Guide: supply-chain trust for AI agents — the tools and skills you didn't write →
Guide: cost and token budgets for autonomous agents — stopping a runaway loop before the bill does →
Guide: designing when an agent should stop and ask a human — escalation without alert fatigue →
Guide: writing task descriptions an agent can actually execute — tickets, issues, and prompts →
Guide: designing persistent memory for agents — what survives across sessions, what doesn't →
Guide: versioning and rolling back agent system prompts — treating instructions like a deployable →
Guide: tracing an agent's decision path — observability beyond print statements →
Guide: rate limits and backoff for agent tool calls — surviving 429s without cascading failure →
Guide: tool contract drift — what happens when the API underneath an agent's tool changes shape →
Guide: checkpointing long-running agent tasks — resuming without redoing or duplicating work →
Guide: deprecating an agent tool without breaking sessions that are already mid-task →
Guide: sizing tool output for agents — pagination, truncation, and summarization at the source →
Guide: caching tool calls for agents — when memoizing a result is safe and when it lies →
Guide: parallel tool calls for agents — which ones are safe to fire together →