skills/florianbruniaux/claude-code-ultimate-guide/source-command-audit-whitepapers/Gen Agent Trust Hub
source-command-audit-whitepapers
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted metadata from repository files to construct shell commands for staleness checks.\n
- Ingestion points: The skill reads
.qmdfiles located in thewhitepapers/fr/,whitepapers/en/,whitepapers/recap-cards/fr/, andwhitepapers/recap-cards/en/directories.\n - Boundary markers: There are no explicit delimiters or instructions provided to the agent to disregard malicious content embedded within the metadata fields of the audited files.\n
- Capability inventory: The skill uses
git logto check commit history and reads multiple directory structures.\n - Sanitization: While the skill mentions parsing the
datefield, it lacks explicit shell-escaping or validation for other interpolated metadata fields, which could be leveraged to manipulate command execution.\n- [COMMAND_EXECUTION]: The skill dynamically constructs shell commands such asgit log --since="{date}"using variables derived from external file frontmatter. If a document's metadata contains shell metacharacters, it could result in unauthorized command execution within the agent's environment.
Audit Metadata