write-release-notes
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local shell scripts (
fetch-release-data.shandvalidate-release-notes.sh) to interact with the GitHub API and validate output length. These scripts use theghCLI and standard utilities likejqandsed. - Evidence:
bash .forge/skills/write-release-notes/scripts/fetch-release-data.sh <version> - Evidence:
echo "<release notes>" | bash .forge/skills/write-release-notes/scripts/validate-release-notes.sh - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from GitHub release bodies and pull request metadata which could contain malicious instructions designed to influence the agent's output during the synthesis phase.
- Ingestion points: The
fetch-release-data.shscript retrieves release bodies and PR titles/descriptions viagh apiandgh pr view. - Boundary markers: The instructions lack explicit delimiters or instructions to ignore embedded commands within the fetched PR data.
- Capability inventory: The agent has access to shell execution (
bash), GitHub CLI (gh), and basic text processing tools. - Sanitization: The fetch script strips ANSI color codes using
sed, but there is no semantic sanitization of the PR content before it is processed by the LLM. - [SAFE]: The skill follows security best practices by not hardcoding credentials, using local authenticated CLI tools, and providing a validation script to ensure output constraints are met.
Audit Metadata