The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs git operations like 'git log', 'git status', and 'git revert', as well as modifications to local project files such as 'plan.md' and 'metadata.json'. These actions are aligned with the skill's primary purpose of managing logical work units.
[SAFE]: The skill implements several robust safety measures, including a requirement for an explicit 'YES' confirmation before execution, the use of non-destructive revert commits instead of history-rewriting resets, and a policy of halting immediately upon encountering merge conflicts or errors.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection through the processing of untrusted data from the local repository. * Ingestion points: The skill reads commit messages using 'git log' and parses content from 'plan.md' and 'conductor/tracks.md' to identify revert targets. * Boundary markers: There are no explicit delimiters or instructions provided to the agent to treat ingested commit messages or file content as untrusted data or to ignore embedded instructions. * Capability inventory: The skill has the capability to execute shell commands ('git revert') and write to the local file system based on the parsed data. * Sanitization: No sanitization or validation of commit messages or track identifiers is specified before they are processed or used in command construction.

conductor-revert