agent-skill-deploy
Pass
Audited by Gen Agent Trust Hub on Mar 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: All included scripts (deploy-preflight.mjs, deploy-analyze.mjs, and deploy-execute.mjs) utilize Node.js
execSyncto execute shell commands for Git and GitHub CLI operations. While version inputs are strictly validated using regular expressions, the use of shell execution for deployment is a high-privilege capability. - [DATA_EXFILTRATION]: The skill is designed to push local repository content, versions, and metadata to external platforms (GitHub and various marketplaces). While this is the intended purpose of the skill, it represents an authorized data transmission capability to external services.
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection in Step 2b of SKILL.md, where the agent is instructed to summarize user-visible changes from
git diffoutput and commit logs. - Ingestion points: Untrusted git commit messages and file diff data processed in SKILL.md (Step 2b) and deploy-analyze.mjs.
- Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore potentially malicious content within the git data.
- Capability inventory: Shell command execution via git/gh and local file system writes.
- Sanitization: No sanitization or filtering is performed on the git history data before it is presented to the agent for analysis.
Audit Metadata