Content
88%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 exceptionally dense, actionable gate-keeping skill: concrete commands with exit-code semantics, a validation-guarded baseline workflow, an explicit triage checklist for failures, and a well-defined coverage/sync procedure. The compressed prose assumes full competence and wastes no tokens on background; the only criticisms are inline time-sensitive operational details and an exhaustive findings catalog that could be split into a reference file.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body never explains concepts Claude already knows — every line carries project-specific fact ("The script calls the laws in packages/plitejs/test/support/recording-laws.ts, the same code the sampled spec calls"), and the compressed style assumes competence. Minor trimmable instances keep it from a 5: time-sensitive specifics such as "Seven recording fixes landed between 2026-10-24 and 2026-11-02" and machine-load timings ("in 28 to 33 s while other sessions load the machine, with load averages from 16 to 77") are not in a deprecated/old-patterns section, and the exhaustive Findings catalog could be tightened. Not score 3, since nothing is padded or unnecessary-to-the-domain. | 4 / 5 |
Actionability | Guidance is fully executable and copy-paste ready: concrete commands ("pnpm check:authored-recording", "pnpm check:authored-recording --lower", "--case='<identity or part of one>'", "--replay='<counterexample>'", "git log --oneline -1 --grep \"check every recording command's accept and reject\""), documented exit codes ("Exit 0 is a pass and exit 1 a finding... Exit 2 is a bad flag or a crashed worker"), and worked examples of identities. This matches the anchor 'Fully executable; copy-paste ready code or commands; specific examples cover the common cases'. | 5 / 5 |
Workflow Clarity | The multi-step process is clearly sequenced with explicit validation checkpoints and feedback loops: run the gate (with the GATE_CASES guard "the run also fails when it compares other than 9,062 cases"), triage a failure against a three-way checklist ("Real bug... Fix the recorder, resolver or history, then run --lower" / "Modeling error... Fix the builder or the script with the reason" / "Intended change... Never weaken a law"), and a sync procedure gated on both a diff review and "a run of pnpm check:authored-recording". Guards like "--lower and --init write nothing when the case count misses GATE_CASES" close the loop on the destructive baseline edit. Matches the anchor 'Clear sequence with explicit validation steps; feedback loops for error recovery; checklists for complex processes'. | 5 / 5 |
Progressive Disclosure | No bundle files exist (references/, scripts/, assets/ are absent), and the body is a self-contained skill whose sections (What it checks, Run it, Read a failure, Findings, What a pass proves, Coverage record, What it does not cover) are well-organized and easy to navigate, with all file pointers being clearly-signaled one-level-deep repo paths. It is not a 5: the exhaustive Findings enumeration (261 gate findings across eight bug classes plus the deep-run catalog) and the mutant analysis are inline content that could plausibly live in a reference file, but they are also the skill's core working knowledge, so the structure is good with minor organization gaps rather than poorly placed. | 4 / 5 |
Total | 18 / 20 Passed |