The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local scripts scripts/chronicle-roll-week.mjs to aggregate git history and scripts/nb-generate.mjs to generate visual assets for the weekly reviews.
[PROMPT_INJECTION]: The instructions include a negative constraint list ('AI-slop refusal list') and strict persona guidelines to enforce a specific tone and prevent characteristic AI phrasing patterns, which is a benign use of instruction-tuning.
[DATA_EXPOSURE]: The skill reads repository-wide git history, changelog data, and project memory files (MEMORY.md) to construct its reflective outputs.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input in the form of git commit messages and changelog entries provided by external contributors or automated systems.
Ingestion points: Git history mined by scripts/chronicle-roll-week.mjs and entries in data/changelog-entries.json.
Boundary markers: Absent; the skill does not specify delimiters or instructions to ignore potential commands embedded in the commit messages.
Capability inventory: The agent has permissions to write files to the documentation directory (docs/chronicle/) and execute project-local scripts.
Sanitization: No sanitization or validation logic is defined for the content extracted from git history before it is summarized into the final report.

starlight-chronicle