The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted metadata, such as package names, version strings, and vulnerability descriptions, from external package registries through commands like npm outdated, pip list, and go list. Maliciously crafted data in a registry could attempt to influence the agent's behavior.
Ingestion points: Audit command outputs used in SKILL.md (Step 1, Step 3, Step 6) and references/upgrade-commands.md.
Boundary markers: None; the retrieved data is processed and presented to the user without explicit delimiters or instructions to ignore embedded commands.
Capability inventory: The skill utilizes shell execution for package managers and test runners (e.g., npm test, pytest, cargo test) throughout its scripts.
Sanitization: The skill does not perform sanitization or validation of external metadata before it is interpolated into the agent's context.
[COMMAND_EXECUTION]: The skill relies on shell command execution to perform dependency auditing, upgrading, and verification.
Evidence: Extensive use of CLI tools including go, npm, pip, cargo, and dotnet in SKILL.md and references/upgrade-commands.md.
[EXTERNAL_DOWNLOADS]: The skill fetches package updates and audit information from well-known, standard package registries.
Evidence: Use of npx npm-check-updates, npm install, pip install, and go get to retrieve remote content.

dep-upgrade