true-spec-worksheet
Skill1 starsChanged 46 days agoMITView on GitHub
Prepare a structured decision worksheet for a requested idea, product behavior, or architecture concern with consequential unresolved choices before formal specification. Use existing authority to avoid reopening settled decisions; do not write final specifications or implementation plans.
In this repository9 other files it gives its agents.
- AGENTS.mdAGENTS.md
- audit-repairSkill
- bug-knowledgeSkill
- documentation-consolidationSkill
- evaluate-repositorySkill
- like-im-5Skill
- project-bootstrapSkill
- spec-driven-developmentSkill
- workflow-orientationSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 1 is distinctive, and 1 is 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.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 10
Before substantial specification, architecture or implementation, demonstrate a concise Understanding Check after relevant inspection and …
- Very specific: line 24
Stop at the worksheet and decision capture.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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
True Specification worksheet
Prepare a useful decision aid for the subject established by the user or conversation. Inspect the repository's instructions, authority hierarchy, index, relevant current product and system specifications, architecture specifications, existing systems map, operational context, and accepted decisions before generating questions. Read only the material relevant to the subject. Missing specifications or a missing map do not force a new project to invent architecture or block independent discovery.
Before substantial specification, architecture or implementation, demonstrate a concise Understanding Check after relevant inspection and before proposing implementation. Keep trivial explicitly authorized work proportional; this is not a new approval gate.
Record settled decisions with their actual sources and owners. Do not reopen them without a changed requirement, demonstrated conflict, or explicit request. Separate governing authority and accepted decisions from facts, implementation evidence, observations, inference, proposals, recommendations, and unresolved questions. Surface conflicting authoritative sources and the owner who must resolve them; do not silently choose a winner.
For each consequential unresolved product or architecture topic, prefer:
Question → recommendation → reasoning → meaningful alternatives → downstream consequences → decision
Order questions by dependency so upstream choices precede dependent ones. Include a recommendation when supported and useful, explaining its assumptions and tradeoffs. Alternatives should expose real choices, not fill a quota. For a consequential architecture choice, explain why the recommendation wins over a credible alternative, its important tradeoff and evidence that would falsify its assumptions or reverse the recommendation. Prototype only material uncertainty under the specification/probe boundary; no multi-model arena or candidate competition is required. Describe consequences for product behavior, ownership, architecture, interfaces, other specifications, or acceptance where they matter. Do not ask questions whose answers follow from existing authority or treat routine implementation choices as new product decisions. Mark genuinely undefined consequential choices for the project's human decision owner. If no consequential choices remain, report that result and identify the specification stage rather than manufacturing questions.
Use the worksheet template when helpful. Finish with decision capture: each choice, its status, owner, acceptance source, affected subjects, dependencies, and remaining gaps. Only explicit acceptance through the project's process can mark a decision accepted; recommendations, preselected options, silence, and worksheet completion do not establish approval.
Return the worksheet in conversation by default. If saving is requested, use the project's provisional design location and naming/version rules, such as its adopted OPS_SPECDEF family; do not impose YMAS on an unrelated repository. The worksheet remains non-authoritative source material even after it records accepted decisions. Keep the actual acceptance sources traceable.
Stop at the worksheet and decision capture. Do not write final SPEC or SPECARC documents, update systems.md, or implement work under this skill. Once the user requests formalization of settled decisions, normal routing can use spec-driven-development; use the development method for the broader lifecycle.
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.

