visual-diff-fix
Pass
Audited by Gen Agent Trust Hub on Jun 15, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill implements a defensive security posture by defining 'Hard Rules' that strictly limit its scope to only the tasks listed in the input JSON and forbid unrelated code refactors or redesigns.
- [COMMAND_EXECUTION]: The skill facilitates the execution of local shell commands for validation purposes, specifically utilizing 'pnpm' scripts defined in the project's 'package.json'. It includes explicit instructions to 'use available scripts only' and 'not invent scripts', which mitigates the risk of arbitrary command injection.
- [INDIRECT_PROMPT_INJECTION]: The skill is subject to indirect prompt injection as it ingests task instructions from external data files.
- Ingestion points: Task data is read from '07_validation/visual_acceptance/fix-tasks.json' and multiple markdown specification files in the 'ui/' directory.
- Boundary markers: The skill uses 'Hard Rules' in 'SKILL.md' and 'Preflight Gates' in 'references/preflight-gates.md' to restrict the agent's behavior and ensure it does not deviate from the provided specifications.
- Capability inventory: The skill possesses the capability to modify source code files in directories like 'src/' and 'app/', and to execute shell commands via 'pnpm' for linting, building, and testing.
- Sanitization: No explicit sanitization or schema validation logic for the content of the tasks is provided within the instructions.
Audit Metadata