adversarial-coach
PassAudited by ClawScan on May 1, 2026.
Overview
This is a coherent instruction-only code review skill; the main thing to notice is that it may prompt the agent to verify code by compiling or running tests.
This skill appears safe for its stated purpose as an adversarial code-review assistant. Before installing or using it, be aware that thorough review may involve compiling or running tests from your project; for untrusted repositories, use a controlled environment and require approval before any command execution.
Findings (1)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
If the agent is allowed to run commands, project build or test commands could execute code from the repository being reviewed.
The skill asks the agent to validate builds, tests, runtime behavior, and functional flows. That may involve running local project commands, which is expected for an implementation-review skill but is still worth user awareness.
| Compilation | Compiles? Tests pass? Runs? | ... | Functional | Test actual flows (not just compilation), verify edge cases work |
Use it in the intended repository, and ask the agent to request confirmation before running build, test, server, migration, or other state-changing commands, especially for untrusted projects.