technical-spec-document
Skill13 starsChanged 5 months agoMITView on GitHub
Translates PRD requirements into an implementation blueprint agreed with the dev team — scope by component, behavior-level contracts, acceptance mapping, estimate structure. Use at dev-estimate time or when the user says "tech spec", "implementation spec", "спецификация", "технический документ". Complements system-design-doc (that one captures constraints; this one specifies what gets built).
In this repository23 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- ambiguity-resolverSkill
- challengeSkill
- deep-thinkSkill
- design-critique-templateSkill
- experiment-designSkill
- ingestSkill
- interview-analysisSkill
- multi-source-signal-synthesiserSkill
- next-advisorSkill
- pipeline-stepsSkill
- post-launch-reviewSkill
- problem-structuringSkill
- product-requirements-docSkill
- setup-initiativeSkill
- solution-scoringSkill
- strategic-narrative-generatorSkill
- system-design-docSkill
- tracking-and-funnelsSkill
- ui-pattern-librarySkill
- user-persona-builderSkill
- user-story-generatorSkill
- user-testingSkill
The file
Files
Technical Spec (implementation blueprint)
Boundary: this skill = what will be built for THIS initiative, agreed
with the dev team. What the system already looks like and what constrains
us is system-design-doc.
The spec's customer is the dev team; its source of truth is PRD §6–8 (which traces to confirmed hypotheses in the registry). Nothing enters the spec without tracing back to a requirement — and no requirement leaves without acceptance criteria.
Structure
- Scope by component — for each affected component (from system-design-doc): what changes, and what explicitly does NOT change (the out-of-scope list prevents estimate creep).
- Contracts at behavior level — new/changed interfaces described as
inputs → outputs → errors. Enough for the dev team to design against;
not pretending to be their design. Example:
submit_order(cart, payment_method) → order_id + confirmation event; errors: payment_declined (retryable), out_of_stock (per item, cart preserved)— behavior, not schemas. - Acceptance mapping — every user story from PRD §8 → which
component(s) deliver it → how we verify (feeds
/ticketsGiven/When/Then directly). - Tracking requirements — events the solution must emit so the AB test (step 14) and post-launch review can measure it. A solution that ships unmeasurable is a spec bug.
- Estimate structure — S/M/L per component with the dev lead's confidence; unknowns from system-design-doc widen ranges explicitly.
Output
output/dev-estimate.md (spec section) + PRD §9–10 updates. Feeds
/tickets — Epics/Stories are generated from the acceptance mapping, so
sloppy mapping here becomes sloppy tickets there.
Anti-patterns
- Duplicating system-design-doc content — link to it instead.
- Code-level detail (schemas, class names) — that's the dev team's design.
- Specs for features whose hypotheses are not confirmed in the registry —
check
hypotheses.py show <dir>before writing a line.
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.

