The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and processes data from the local repository that may be influenced by external contributors.
Ingestion points: Data is ingested from README.md, CHANGELOG.md, and git commit history (via git log) in SKILL.md.
Boundary markers: The sub-agent prompts (Agent A and Agent C) do not include explicit delimiters or instructions to ignore embedded commands within the ingested text.
Capability inventory: The skill has broad capabilities including file reading (Read), file writing (Write, Edit), and shell command execution (Task).
Sanitization: There is no evidence of sanitization or filtering of the repository content before it is processed by the LLM sub-agents.
[COMMAND_EXECUTION]: The skill utilizes the Task tool to execute shell commands for analyzing repository history.
Evidence: In SKILL.md, Step 3 (Agent C) executes git log commands to determine when the README was last updated and to extract commit messages since that date.

update-readme