gs-why
Pass
Audited by Gen Agent Trust Hub on Nov 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from multiple external sources which may contain embedded instructions designed to influence the agent's behavior.
- Ingestion points: Data is ingested from Git commit history, Pull Request descriptions and comments, issue trackers (Linear, Jira), long-form documents (Notion, Confluence), team chat (Slack), and observability platforms (Datadog, Sentry).
- Boundary markers: The prompt templates in
investigator-prompt.mdandsynthesizer-prompt.mduse markdown headers and sections like> {QUESTION}to separate instructions from data, but they lack explicit warnings for the model to ignore embedded instructions within the retrieved content. - Capability inventory: The skill possesses capabilities to execute shell commands and call various MCP-based tools to fetch and query data.
- Sanitization: There is no evidence of explicit sanitization or filtering of the content retrieved from external sources before it is processed by the agent.
- [COMMAND_EXECUTION]: The skill generates and executes shell commands to interact with the project's version control and host platform.
- Evidence: Instructions in
SKILL.mdandreferences/sources/code-archaeology.mddirect the agent to run commands such asgit blame,git log,gh pr view, andglab mr viewusing parameters derived from the conversation or repository metadata. - Context: These commands are fundamental to the skill's purpose and use well-known CLI tools to retrieve information from legitimate service providers.
Audit Metadata