The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes several shell commands to gather project metadata, including git rev-parse, git branch, git log, git worktree list, git status, gh pr list, and gh issue list. These are standard diagnostic commands used to build the handover context.
[DATA_EXFILTRATION]: The skill collects sensitive repository information, including branch names, recent logs, worktree paths, and GitHub Issue/PR details. While this data is intended for a local handover document (docs/handover/), the aggregation of this metadata represents a data exposure surface if the output is shared externally.
[PROMPT_INJECTION]: The skill is susceptible to Indirect Prompt Injection (Category 8) because it ingests untrusted data from external sources without explicit sanitization or boundary markers.
Ingestion points: Data enters the context via git log, gh pr list, and gh issue list (SKILL.md, Procedure Step 1).
Boundary markers: The instructions do not specify the use of delimiters (e.g., XML tags or triple backticks) to isolate the collected data from the agent's instructions during document generation.
Capability inventory: The skill has the capability to execute shell commands and write to the local file system (SKILL.md, Procedure Step 5).
Sanitization: There is no mention of escaping, filtering, or validating the content retrieved from commit messages, PR titles, or issue labels.

handover