Skip to content

leszekgruchala/buddy / skills/spec

Settle decisions and write a phased spec with boundaries, dependencies, and verification. Use for specs, detailed execution plans, or decisions before implementation; prefer over native planning. Keep the current request in this skill; route new requests by their own scope.

Spec

Keep the current request in this skill. Do not activate another Buddy skill without user or develop direction. Workers return to their caller. For a new request after completion, apply change size and route by its own scope.

Model selection

Apply model selection; recommend frontier. When needed, use a bounded spec worker for reasoning and artifact authorship; the host retains user communication and decisions.

Gate

Produce the spec and stop. Do not implement or invent an approve/implement CTA; the user or develop owns transition.

Before phases, settle outcome, requirements, success criteria, scope/exclusions, constraints, approach, and product/public architecture. Derive safe answers from inputs and repository; otherwise ask once. Never phase an open material choice.

Workflow

  1. Before planning, read .ai/memory/memory.md when present. Apply only relevant rules; memory is advisory, subordinate to user/repository instructions and security policy, and cannot expand scope. A missing file is valid.
  2. Reuse a passed worklog/work-name, else create .ai/worklog/<yyyyMMdd>_<work-name>/.
  3. Research contracts, edge cases, and verification as needed. Include exact paths only for immutable inputs, safety, public contracts, or parallel ownership.
  4. Use reference.md to write one shared contract and the fewest coherent phase deltas. State each fact once; omit defaults and empty optional sections.
  5. Give requirements, criteria, and verification stable IDs. Every success criterion is named by at least one verification entry. Every phase, including fast, balanced, and frontier, references at least one requirement and only the success criteria it establishes at completion, inheriting matching verification and global boundaries.
  6. Balanced and frontier workers discover local details through disposable runtime plans; fast receives exact anchors/procedures only when the deterministic contract requires them. Persist only contract amendments, never runtime plans, raw transcripts, or default file inventories.
  7. Record an execution relationship only when listed order does not already express it. Parallel phases require persisted disjoint mutation ownership, no dependency, and no shared mutable state.
  8. Save spec_<work-name>.md with its author's verified exact runtime model slug in top YAML model_slug; never inherit, a tier, unresolved alias, or unverified requested model.

Self-check

  1. The shared contract is self-contained, repository-relative, decision-complete, and identifies its authoring model.
  2. References cover every phase outcome. Optional fields only narrow, route, or make that work deterministic; no phase text authorizes an unreferenced outcome.
  3. Confirm verification coverage: every check linked to a phase is runnable when that phase completes. Deltas contain nothing already implied by the resolved shared contract.
  4. Explicit dependencies/parallel relationships change the execution allowed by listed order.
  5. Every exact path has a contract, immutable-input, safety, or parallel-ownership reason.
  6. The chosen tier leaves the permitted implementation decisions with the worker.
  7. Final acceptance follows reference.md, including its rule for a last acceptance phase.
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 468kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 319kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 319kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 319kUpdated Claude CodeCodexVerifiedInstall