docs-workflow
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from GitHub Pull Requests (titles, descriptions, and file changes) to perform documentation analysis. This external data represents a surface for indirect prompt injection.
- Ingestion points: External PR content is queried from the grafana/tempo repository using the GitHub CLI as shown in the eval commands.
- Boundary markers: The workflow does not explicitly define delimiters or specific instructions to isolate and ignore embedded prompts within the ingested PR data.
- Capability inventory: The skill uses the Bash, Read, Grep, and Write tools to modify documentation and interact with the local repository.
- Sanitization: There is no evidence of specific sanitization or validation of the PR content before it is processed by the agent.
- [COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute repository commands and manage the documentation pipeline.
- [SAFE]: All external interactions are restricted to official vendor repositories and tools, which is consistent with the skill's intended purpose and the author's identity.
Audit Metadata