The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches release metadata and archives from GitHub using the gh CLI and the npx skills command. These operations are targeted at the vendor's own repository (wilsonkichoi/skills-manage-validation-20261002).
[REMOTE_CODE_EXECUTION]: The skill uses npx skills@1.7.0 to download and install packages directly from a remote GitHub repository using Git URLs with tags (e.g., https://github.com/wilsonkichoi/skills-manage-validation-20261002.git#${tag}).
[COMMAND_EXECUTION]: The skill requires the execution of multiple shell commands, including git status, git archive, gh api, and npx, which are used for environment inspection, evidence verification, and project mutation.
[DYNAMIC_EXECUTION]: Installation commands are dynamically constructed using variables for the repository identity and resolved version tags, which are then passed to npx for execution.
[INDIRECT_PROMPT_INJECTION]: The skill processes natural language content from external sources, specifically README files and migration guidance contained within remote GitHub release archives. This introduces an attack surface where instructions embedded in those files could influence agent behavior.
Ingestion points: Remote Git archives (README, migration guidance) and GitHub API release metadata.
Boundary markers: The skill does not define explicit delimiters or 'ignore' instructions for the content read from remote archives.
Capability inventory: File system modification (via git and npx), network access (via gh and npx), and package installation.
Sanitization: While the skill performs schema validation on the skills-lock.json file, it lacks explicit sanitization for natural language instructions found in the documentation it fetches.

wkc-manage