The Agent Skills Directory

[COMMAND_EXECUTION]: The deployment workflow in SKILL.md instructs the agent to execute a shell command echo '{{CHANGELOG}}' > /tmp/release-notes.md where {{CHANGELOG}} is text derived from external git commit messages. This pattern is vulnerable to command injection if the content contains single quotes that escape the shell string context.
[COMMAND_EXECUTION]: The skill utilizes the GitHub CLI (gh) to automate the creation of repository releases and handles git operations such as tagging and pushing to remote repositories.
[INDIRECT_PROMPT_INJECTION]: The skill has a vulnerability surface where untrusted data from the repository environment can influence agent actions.
Ingestion points: The scripts/deploy-analyze.mjs script retrieves git commit logs and file diffs which are subsequently summarized by the agent.
Boundary markers: The instructions do not provide explicit delimiters or "ignore embedded instructions" warnings for the processed git data.
Capability inventory: The skill uses the Bash tool to perform sensitive operations, including repository management and release publishing.
Sanitization: No sanitization or shell-escaping is performed on the commit message content before it is interpolated into shell commands or processed by the agent.

agent-skill-deploy