Content
92%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 orchestration-style SKILL.md: lean, sequentially clear with explicit validation and feedback loops, and a well-structured one-level-deep reference bundle where every cited file and anchor exists. The body intentionally delegates executable detail to its references, which is appropriate progressive disclosure and only mildly limits standalone actionability.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The ~45-line body is dense with task- and product-specific information — a decision table, terse per-step directives, and an constraints list — with no explanation of concepts Claude already knows and no padding. It assumes competence ("Inspect installed package source when documentation does not match the installed API") and every sentence carries non-obvious information, matching the lean score-5 anchor rather than the minor-trimming of score 4. | 5 / 5 |
Actionability | Concrete, executable elements are present: the command pattern "iris agent-playbooks <command> --help", the hard rule "Every `iris` command uses root options `--json --no-input`", specific directives ("Register disclosure tools with both the model and executor", "Keep runtime `.env` credentials separate from `.env.content`"), and purposeful pointers to per-step reference files. It stops short of a 5 because the body itself contains no copy-paste-ready code or command for the integration step — the executable detail is delegated to references — leaving minor gaps if the body is read alone. | 4 / 5 |
Workflow Clarity | Six numbered phases (Orient → Prerequisites → Publish → Integrate → Verify → Hand off) give a clear sequence with explicit checkpoints: "Confirm its serving version before wiring the agent", and in section 5 a full feedback loop — "inspect the content keys/versions and procedure loads, check a paraphrase and missing-content behavior, fix observed gaps, and rerun". Destructive cleanup is guarded ("run affected application checks on the final integration"), matching the score-5 anchor with validation steps and error-recovery loops. | 5 / 5 |
Progressive Disclosure | The body is a concise overview in which every one of the six bundle files in references/ is cited with its purpose (e.g., "Follow [provisioning](references/provisioning.md) for the short prerequisite check"), conditional paths are clearly signaled ("For a semantic response cache, read [cache migration](references/response-cache.md) before..."), both anchor links resolve to real headers (## Handoff, ## Maintain the content), and references are one level deep with no nested chains. This matches the score-5 anchor of a clear overview with well-signaled one-level-deep references. | 5 / 5 |
Total | 19 / 20 Passed |