marketplace/bundles/plan-marshall/skills/persona-code-reviewer/SKILL.md
Code-review persona that composes other personas as evaluation lenses
npx skillsauth add cuioss/plan-marshall persona-code-reviewerInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
REFERENCE MODE: This skill is a persona shell. It declares the code-reviewer evaluator identity and the composition it resolves to; it carries no executable workflow of its own.
The code reviewer is a meta/evaluator persona — it has no work-activity profile of its own. Instead it composes other personas as evaluation lenses, reading a change through each composed persona's standards to judge correctness and intent. Because it owns no primary work-activity profile, it omits the profiles: field entirely.
The persona resolver (manage-personas resolve) flattens this persona's composition DAG into a deduped skills[]:
plan-marshall:persona-plan-marshall-agent (the unconditional foundational base every persona inherits).composes: frontmatter (persona-implementer, persona-module-tester, persona-integration-tester, persona-documenter, persona-security-expert); the resolver recurses into each composed persona's refs and profiles, then dedups.None. A meta/evaluator persona that composes other personas as lenses omits the profiles: field — it is never reverse-looked-up from a task's work-activity profile.
This persona provides the quality verify profile for the ext-point-verify findings-pipeline stage, mirroring the role persona-security-expert plays for the security profile. The adversarial-refute methodology lives in standards/adversarial-refute.md: when a producer declares metadata.verification_profile: quality, the orchestrator's verify pre-stage resolves that profile key to this persona and loads that standard in-context to refute each candidate quality/structural/documentation finding before triage. Confirmed findings flow on to ext-triage-* unchanged; refuted false positives close rejected and never reach triage. The stage placement and rejected semantics are owned by the ext-point-verify contract — see that document.
standards/adversarial-refute.md — The quality verify-profile methodology this persona provides, implementing ext-point-verify as a verify profile (the quality counterpart to the security persona's role)extension-api:ext-point-verify.md — The verify extension-point contract the quality profile implementsdevelopment
Domain-owned OpenRewrite log-line finding parser for the java-cui domain — parses the
development
Domain-owned OpenRewrite marker detection for the java-cui domain — scans Java/Kotlin sources for cui-rewrite TODO markers, categorizes them by recipe, and fails the gate on any detected marker
development
Operator control surface for the marshalld build server — enrol/drop a project in the machine-global registry (the opt-in enable signal and anti-laundering wall), manage the daemon lifecycle (start, stop, drain, status, install, upgrade) version-pinned to the verified bundle copy, and inspect the daemon's per-project interaction-audit log (read-only)
tools
The tiny build-consumption client for the marshalld build server — submit a build job, bounded long-poll for its result, ping the daemon identity, and preflight registry-plus-liveness in one call; consumption only, never provisioning or enrolment