Content
82%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.
A strong, highly actionable reference: executable commands with concrete examples, an explicit failure-mode recovery table, and tight information density. Its weaknesses are modest — no explicit step-by-step workflow framing, a dangling docs/frida.md reference unverifiable against the (empty) bundle, and a few verbose passages.
Suggestions
Move the "Evidence caveat" paragraph and the pipeline-integration consumer table into a separate reference file (e.g. references/pipeline.md) and link to it, keeping SKILL.md as a lean overview.
Condense the "Status" section to a few lines; roadmap details (planned integrations, abandoned PR #57) dilute the operational focus.
Add a brief numbered "basic flow" (install → verify with raptor doctor → invoke → read metadata.json → consult failure modes) to make the implicit workflow sequence explicit.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | A dense, information-rich reference with no explanations of concepts Claude already knows, but several sentences carry project-roadmap detail (abandoned PR #57, planned integrations) and the "Evidence caveat" paragraph could be tightened, so anchor 4 fits better than 5. | 4 / 5 |
Actionability | Fully executable guidance throughout: install commands, a complete CLI invocation with all flags, five commented example commands covering local/spawn/USB/remote/custom-script cases, and a copy-paste-ready Python API block with three concrete variants. | 5 / 5 |
Workflow Clarity | Sequence is legible via ordered sections and includes explicit validation/recovery hooks: `raptor doctor` confirms the install, and the "Failure modes (read metadata.json first)" table maps concrete error fragments to causes and fixes. Scored 4 rather than 5 because the steps are implicit in section order rather than explicitly enumerated as a workflow. | 4 / 5 |
Progressive Disclosure | Well-organized single-file skill with clear section headers and one clearly signaled one-level reference ("See `docs/frida.md"); no bundle files exist to offload detail into, and dense sections like the threat model and pipeline-integration table are borderline candidates for extraction. Anchor 4 fits: good structure with minor organization gaps. | 4 / 5 |
Total | 17 / 20 Passed |