The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes external data from package changelogs which could theoretically contain malicious instructions.
Ingestion points: The script references/check-outdated.ts fetches package release notes and bodies from the GitHub API (api.github.com).
Boundary markers: None. The raw release note body is stored in outdated-changelog.json and interpreted by the agent for feature adoption.
Capability inventory: The skill has capabilities to write files, execute shell commands via bunx (references/skills-update-local.ts), and make network requests.
Sanitization: No sanitization or filtering is performed on the external content fetched from GitHub before it is processed.
[EXTERNAL_DOWNLOADS]: The skill fetches data from trusted external registries and services as part of its core functionality.
references/check-outdated.ts makes requests to registry.npmjs.org and api.github.com to check for updates and retrieve release history.
[COMMAND_EXECUTION]: The skill executes shell commands to update other agent skills installed in the project.
references/skills-update-local.ts uses Bun.spawn to execute bunx skills@latest add for sources defined in the local skills-lock.json file.

update-packages