The Agent Skills Directory

[COMMAND_EXECUTION]: The skill defines a recursive model for executing various Git commands to manage repository state, including git add ., git commit, git pull --no-rebase, git push, and git merge. These operations are applied to both the parent repository and its submodules.
[DATA_EXFILTRATION]: The skill includes functionality to automatically push all local changes to upstream remotes using git push. This is a network operation that transmits repository content externally.
[PROMPT_INJECTION]: The skill exposes an indirect prompt injection surface by instructing the agent to read and act upon untrusted content from the repository (e.g., during conflict resolution or commit message generation).
Ingestion points: The agent is instructed to use git status, git diff, git ls-files, git show, and git log to inspect file content and metadata (SKILL.md).
Boundary markers: The skill lacks explicit boundary markers or instructions to isolate the ingested data from the agent's control logic.
Capability inventory: The agent can execute file system modifications (git add, git commit) and network synchronization (git push, git pull) based on its processing of the repository data (SKILL.md).
Sanitization: There is no evidence of content sanitization or validation of the data retrieved from the repository before it is presented to the model.

sync