task-spec
Skill0 starsChanged 7 days agoMITView on GitHub
Writes a clear task specification before delegating work to a coding agent or starting a non-trivial change. Use when the user describes a feature, bug, refactor or automation loosely; before long-running or unattended agent work; or when the user asks for a spec, plan, ticket, goal or "done when" criteria. Produces Goal / Context / Constraints / Done when, with a verifiable check.
- Reads credentials
5 other projects use a version of this.
Adopted by
Most recently taken up by
- emersonnakashima/task-spectheir own version, first seen Nov 15, 2026
- hammoda711/task-spec-skilltheir own version, first seen Nov 8, 2026
- eyeix/skillstheir own version, first seen Oct 13, 2026
- MagicianDu/task-spec-studiotheir own version, first seen Jul 1, 2026
- YorkWong1995/OPC-Harnesstheir own version, first seen May 26, 2026
Other versions5 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 5
- YorkWong1995/OPC-Harness★ 4same name
- eyeix/skills★ 1same name
- emersonnakashima/task-spec★ 0same name
- hammoda711/task-spec-skill★ 0same name
- MagicianDu/task-spec-studio★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository11 other files it gives its agents.
- drop-in/AGENTS.mdAGENTS.md
- llms.txtllms.txt
- clarify-firstSkill
- coverage-checkSkill
- fresh-eyes-reviewSkill
- fresh-start-handoffSkill
- grounded-answerSkill
- honest-critiqueSkill
- learning-tutorSkill
- prompt-improverSkill
- verify-claimsSkill
The file
We split this skill into the rules it gives an agent
2 rules. 0 are common, 1 is distinctive, and 1 is very specific to this project.
0
Common
Most of the 3 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 35
Make "done" verifiable: Every item under *Done when* should be checkable by running something, not by opinion.
- Very specific: line 37
Move non-negotiable rules out of prose: If a rule must never be broken (don't read .env, don't touch migrations/), recommend enforcing it …
“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Task spec
Most disappointing agent sessions trace back to missing information, no way for the agent to check its own work, or rules nothing enforces. A short spec fixes the first two.
Steps
- Gather facts, but don't stall. If you can read the code, docs or data, read the relevant parts first so the spec refers to real files and behavior (don't change anything yet). If you can't access them, or the user only wants the instructions written, write the spec now from what you were told: put unknowns in angle brackets (
<file?>,<current p95 latency?>) and list the questions or measurements needed to fill them. Write the spec:
## Goal What changes and why, in one or two sentences. ## Context - Relevant files: ... - Current behavior / symptom (exact error, reproduction steps): ... - Existing pattern to follow: ... ## Constraints - Out of scope / don't touch: ... - No new dependencies without asking. ## Done when - [ ] A check that can be run and returns pass/fail (test, build, script, screenshot comparison) - [ ] ... - [ ] Report: what changed, the commands run and their output, remaining risks- Make "done" verifiable. Every item under Done when should be checkable by running something, not by opinion. For bugs, the first item is a failing test that reproduces the bug (red), which must pass afterwards (green).
- Size the process to the task. A small bug needs a paragraph; a large feature may need a plan the user approves before any code is written.
- Move non-negotiable rules out of prose. If a rule must never be broken (don't read
.env, don't touchmigrations/), recommend enforcing it with permission rules, hooks or CI rather than the spec alone.
Evidence
- Goal / Context / Constraints / Done when: OpenAI Codex best practices — https://learn.chatgpt.com/guides/best-practices
- Four causes of disappointing sessions; give the agent a check it can run; deterministic vs non-deterministic controls: Claude Code best practices — https://code.claude.com/docs/en/best-practices
- Goals should state outcome, constraints and verification: OpenAI Codex long-running work — https://learn.chatgpt.com/docs/long-running-work
- Red/green TDD with agents: Simon Willison — https://simonwillison.net/guides/agentic-engineering-patterns/red-green-tdd/
- More in the guide: https://github.com/ji-hun-git/how-to-use-ai/blob/main/docs/07-ai-coding.md
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

