daily-docs-fold
Pass
Audited by Gen Agent Trust Hub on Jun 28, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a bundled Python script (
scripts/check-doc-indexes.py) to perform documentation linting, routing checks, and token budget estimations within the target repository. - [DATA_EXFILTRATION]: The skill is designed to report repository-specific evidence and worklog content to external GitHub issue trackers when identifying bugs in shared skills. While GitHub is a trusted service, the process involves transmitting local research and telemetry data to an external platform.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. It is instructed to ingest data from untrusted ingestion points (such as
ephemeral/andraw-material/directories) and 'fold' that content directly into core agent policy files likeAGENTS.mdorCLAUDE.mdwhich define the agent's behavior and constraints. - Ingestion points: Files located in
ephemeral/,raw-material/, worklogs, research notes, and user corrections. - Boundary markers: Absent. The instructions do not specify the use of delimiters or warnings to prevent the agent from obeying instructions embedded in the ingested text.
- Capability inventory: The skill can modify sensitive configuration files (
AGENTS.md,CLAUDE.md), execute local scripts (check-doc-indexes.py), and perform network operations (GitHub issue creation). - Sanitization: Absent. There is no mention of filtering or escaping content before it is incorporated into the durable documentation.
Audit Metadata