hassan-mohiddin/freeflow / .deprecated/skill-tooling-pre-rewrite/skills/evaluate-skill
evaluate-skill
VerifiedUse when evaluating, comparing, or making readiness claims about agent-skill behavior.
references/token-efficient-execution.md
View on GitHub# Token-Efficient Execution
Optimize provider work without adding caller-managed lifecycle machinery.
## Order
1. Run deterministic preflight before any provider request.
2. Start with one strong pressure case.
3. Run objective graders before semantic graders.
4. Invoke semantic grading only for unresolved fixed assertions.
5. Inspect result summaries before raw transcripts.
6. Rerun the whole case only after a measured change or infrastructure diagnosis.
7. Expand cases, models, or hosts only when the support claim requires them.
## Planning
Before execution, report:
- one selected case and its ordered variants;
- host, evidence classes, tools, and isolation policy;
- model and thinking settings when applicable;
- maximum subject and semantic Pi-process counts;
- per-process turn, timeout, and output limits;
- worst-case approved turns;
- spend ceiling or unavailable-cost limitation;
- unsupported or reduced-fidelity evidence.
Bootstrap runs variants serially. It has no cache, batching, concurrency, adaptive repeats, resume, or partial reuse. These mechanisms save work only by adding lifecycle state the caller or runtime must trust.
A hard-limit or infrastructure failure publishes diagnostics, not a gradeable result. The output limit applies to retained canonical evidence after cumulative Pi JSON update snapshots are compacted; a separate internal safeguard bounds raw transport. After the cause and owner-approved limits are resolved, invoke the whole case again.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| 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: 471k | Updated | Claude CodeCodex | Verified | Install |
| 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: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| 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 available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |