Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/michaelcjoseph/agent-coding-setup/worknpx skills add michaelcjoseph/agent-coding-setup --skill workgit clone --depth 1 https://github.com/michaelcjoseph/agent-coding-setupWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00000 | $0.04063 |
| Opus 5 | $0.00000 | $0.02031 |
| Sonnet 5 | $0.00000 | $0.00813 |
| Haiku 4.5 | $0.00000 | $0.00406 |
Grade A, and why
work scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 347 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/work
Pick the next pending task from a project's task list and drive it through the full implementation cycle: plan, implement, test, review, fix, simplify, and mark complete.
Usage
/work [project-name] [--auto]
If no argument is given, list all projects in docs/projects/ and ask the user which one to work on.
--auto (unattended mode): skips all user-interaction gates so one invocation sweeps the entire task list end-to-end. Specifically: step 2 skips the "confirm before proceeding" gate, Phase 1 skips Plan Mode entirely (still does exploration and writes the plan to the turn output for the transcript record, but does not call EnterPlanMode/ExitPlanMode since those require human approval), step 15 creates one git commit per completed task (using the release-notes agent to draft the message) without pausing for approval, and step 16 auto-continues to the next unchecked task without asking. Hard stops (step 2 empty list, step 9 caused-by-diff failures unresolved, step 11 BLOCK verdict, step 12 tests failing twice, step 15 commit still failing after one retry, step 16 empty list) still terminate the run — those are error/completion exits, not checkpoints.
Because --auto commits everything in the working tree at the end of each task, start from a clean working tree (no unrelated uncommitted changes). If git status is dirty at step 2 in --auto mode, stop and report to the user instead of sweeping those changes into a task commit.
No discretionary pauses in --auto. The hard stops listed above are the only valid reasons to stop or ask the user something mid-run. Do not invent additional checkpoints — do not pause to offer a commit, confirm direction, summarize progress, checkpoint a subsystem boundary, or ask whether to proceed between tasks. If a general instruction elsewhere (e.g., "only commit when the user asks" or "confirm risky actions") seems to conflict with --auto, resolve it by not performing the action, not by pausing. The user authorized an unattended sweep by passing --auto; interrupting that sweep for any reason outside the hard-stop list violates the contract.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 347 lines · 0 tokens per session scan A 1f109cb04c6d
work is a skill published in the GitHub repository michaelcjoseph/agent-coding-setup (2 stars, last pushed 4mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 4,063 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
babysit-pr
Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…
imagegen
Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
next-cache-components-optimizer
Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…