Content
100%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
An exemplary instruction-only skill body: terse, every step executable with exact strings/paths/commands, an explicit validation-before-push checkpoint, a self-limiting budget mechanism with a stated rationale, and clear escalation boundaries. Nothing is over-explained and nothing essential is missing.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The 22-line body is lean and operational with no padding and no explanation of concepts Claude already knows; even the one rationale sentence ("Tessl has no run budget of its own, so this count is the only thing that stops a loop between review and fix") earns its place by justifying the attempt cap. Every token is load-bearing. | 5 / 5 |
Actionability | Guidance is fully concrete: the exact marker "<!-- perihelion-factory:shepherd -->", the exact label "factory:needs-human", the exact input path ".cloud-launch/trigger-events.json", and specific commands ("Run the build, the tests and `tessl change verify`"). As an instruction-only skill, its guidance is copy-paste-actionable throughout, matching the top anchor. | 5 / 5 |
Workflow Clarity | Six clearly sequenced steps with an explicit validation checkpoint (build/tests/verify before commit and push, step 5), a feedback loop for failed checks (read log, reproduce locally, fix the cause, step 4), and an upfront budget check plus enumerated stop-and-hand-back conditions. This matches the top anchor: clear sequence, explicit validation, error-recovery guidance. | 5 / 5 |
Progressive Disclosure | At under 50 lines with no bundle files needed, it is cleanly sectioned (Inputs, Procedure, Stop and hand back) and delegates detail via a single well-signaled one-level reference to the plugin's `resolve-feedback` skill. Per the under-50-line guideline, this merits the top score. | 5 / 5 |
Total | 20 / 20 Passed |