goal-crafter
Skill192 starsChanged 22 days agoMITView on GitHub
Draft a verifiable goal prompt from an idea or task. Use when the user asks for a goal, completion criteria, or a prompt to hand to another agent; this skill writes the goal rather than executing or scheduling it.
In this repository43 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- ask-mattSkill
- codebase-designSkill
- code-reviewSkill
- diagnosing-bugsSkill
- implementSkill
- implement-specSkill
- prototypeSkill
- researchSkill
- resolving-merge-conflictsSkill
- tddSkill
- domain-modelingSkill
- execute-spec-in-forkSkill
- grill-with-docsSkill
- improve-codebase-architectureSkill
- prSkill
- retroSkill
- roundtableSkill
- setup-matt-pocock-skillsSkill
- spec-executorSkill
- to-goalSkill
- to-specSkill
- to-ticketsSkill
- triageSkill
- wayfinderSkill
- wizardSkill
- claude-handoffSkill
- loop-meSkill
- setup-ts-deep-modulesSkill
- writing-beatsSkill
- writing-fragmentsSkill
- writing-shapeSkill
- git-guardrails-claude-codeSkill
- migrate-to-shoehornSkill
- scaffold-exercisesSkill
- setup-pre-commitSkill
- grillingSkill
- grill-meSkill
- handoffSkill
- teachSkill
- to-questionnaireSkill
- wait-whatSkill
- writing-for-agentsSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 10
Use the conversation and supplied evidence first.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Goal Crafter
Turn the user's intended outcome into a concise goal that another agent can execute and verify.
Use the conversation and supplied evidence first. Ask only about missing decisions that change the scope, completion criteria, or permissions. Already-approved material needs no new interview; if it is missing an essential decision, identify the gap rather than inventing an answer.
A useful goal states:
- Outcome: the result the user wants, with the relevant workspace or target.
- Completion: observable evidence that distinguishes finished from unfinished work. Preserve the user's criteria; do not substitute green tests for product acceptance. If final acceptance needs the user, name that boundary.
- Constraints: material scope, authorization, budget, or files to protect, when applicable.
- Context: the sources and starting points the executor actually needs.
Use the shortest format that carries those facts. Adapt to a named harness when its format is known; otherwise use plain Markdown. A small task may need only a few sentences. Include an execution order only when dependencies matter. Do not prescribe a fixed interview, number of checkboxes, or engineering method.
Preserve authority from the source: writing a goal does not authorize implementation, commit, push, deployment, messages, or scheduling. Do not promise an unattended run will necessarily finish. Keep explicitly requested capability or model choices; omit generic model-tier recommendations.
Deliver the goal ready to copy, plus only essential unresolved questions. This is goal authoring. to-goal separately compiles an approved source against repository state, and spec-executor performs implementation; neither is required to use this skill.
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.

