The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses standard CLI tools such as find, grep, and git, as well as language-specific test runners like cargo test and npm test. It also invokes a local script plugins/sanctum/scripts/update_versions.py. These operations are transparently described and serve the skill's primary purpose of version management.
[INDIRECT_PROMPT_INJECTION]: The skill reads data from local configuration and documentation files, creating a potential surface for indirect prompt injection.
Ingestion points: The skill reads Cargo.toml, package.json, pyproject.toml, CHANGELOG.md, and various markdown files in the docs/ and book/ directories.
Boundary markers: There are no explicit delimiters or boundary markers specified to separate file content from the agent's instructions.
Capability inventory: The skill can modify local files and execute shell commands, including build tools and test runners.
Sanitization: No sanitization or validation of the file content is described before it is processed by the agent.
Context: Given that these operations are standard for development tools and restricted to the local workspace, the risk is consistent with the intended functionality.

version-updates