The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the glab (GitLab CLI) and git tools to perform operations on a self-hosted GitLab instance (https://gitlab-erp-pas.dedalus.lan). These operations include viewing merge requests, fetching diffs, and listing CI pipelines, which are necessary for the skill's stated purpose.
[PROMPT_INJECTION]: The skill has a surface for indirect prompt injection where malicious instructions could be embedded in the data being processed.
Ingestion points: The agent fetches and processes untrusted data such as merge request titles, descriptions, and code diffs using glab mr view and glab mr diff (documented in SKILL.md).
Boundary markers: The instructions lack explicit boundary markers or directives to the agent to ignore or isolate potentially malicious instructions contained within the fetched merge request content.
Capability inventory: The skill has the capability to write back to the GitLab instance using glab mr note and the GitLab API to post discussions (documented in SKILL.md).
Sanitization: There is no mention of sanitizing or validating the content retrieved from GitLab before the agent interprets it.

code-review