using-skillpack-maintenance
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run a repository-local script (
python scripts/check_marketplace_integrity.py) to verify the consistency of the marketplace catalog and its components. This is a standard and expected development workflow for maintenance tasks. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external data from third-party plugins, marketplace manifests (e.g.,
.claude-plugin/marketplace.json), and skill files. Maliciously crafted content within these external resources could attempt to influence the agent's behavior or provide misleading instructions during the review and maintenance process. - Ingestion points: Plugin files under
plugins/, marketplace metadata in.claude-plugin/marketplace.json, and other repository manifests. - Boundary markers: The instructions do not define specific delimiters or "ignore embedded instructions" warnings to isolate external data from the agent's core task instructions.
- Capability inventory: The agent is authorized to modify repository files ("implement fixes") and execute specific local validation scripts.
- Sanitization: There are no explicit instructions for sanitizing or escaping the content of the plugins being reviewed before the agent processes them.
Audit Metadata