Conciseness

The ~38-line body is lean and assumes competence: every line is a rule, a command, an exit code, or a done-condition, with zero filler and no explanation of what mutation testing is. This matches the 'lean and efficient; every token earns its place' anchor; it is above anchor 4 because there is nothing to trim.

Actionability

Mostly executable guidance: a copy-paste-ready waive command with argument syntax ("bob waive <file::function|Target> --reason \"texto de uma linha\""), exit codes 3 and 4, and the report path .bob/report.md. It stays at anchor 4 rather than 5 because the primary path — step 2, writing the killing test — gives a precise criterion but no concrete example of such a test.

Workflow Clarity

The numbered steps each close with an explicit validation checkpoint ("Feito quando uma nova bob mutate não lista mais esse sobrevivente sem waiver"; "Feito quando o waiver aparece no topo de .bob/report.md e o exit deixa de ser 3"), include a repeat loop ("Repetir até não restar sobrevivente sem waiver"), and handle error recovery (waive exiting 4 on an invalid target, the '—' target rule). This matches the anchor for clear sequence with explicit validation, feedback loops, and error recovery.

Progressive Disclosure

The skill is under 50 lines, single-purpose, and organized into clear sections (Quando, Passos, Dentro do razoável, Nunca enfraquecer o código de produção) with no bundle files present or needed. Per the rubric's simple-skill exception, this warrants a 5 with just well-organized sections.

Specificity

The description names the domain and two concrete actions — "Matar um mutante sobrevivente" and "gravar um waiver" — but stops short of detailing what killing a mutant concretely involves (writing a stronger test), matching the '1-2 concrete actions, not comprehensive' anchor. It is above anchor 2 because the actions are domain-specific rather than generic.

Completeness

It explicitly answers both what ("Matar um mutante sobrevivente dentro do razoável, ou gravar um waiver") and when via an explicit "Use quando" clause with two concrete trigger conditions (exit code 3 from specific commands; the report listing a survivor without waiver). This matches the anchor requiring both what AND when with concrete trigger phrases.

Trigger Term Quality

It includes the exact natural trigger phrases a user of this tool would say: the verbatim commands "`bob mutate`" and "`bob check --gates mutate`", "sair com 3", "sobrevivente", "waiver", and "relatório". A few natural synonyms (e.g., 'mutation testing', 'surviving mutant') are missing, keeping it at anchor 4 rather than 5.

Distinctiveness Conflict Risk

The description is anchored to a clear niche with distinct triggers: tool-specific commands, an exact exit code, and a specific report condition, so it is unlikely to trigger for the wrong skill. This fits the 'clear niche with distinct triggers; minimal conflict risk' anchor.