The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute local Python scripts (pm-schema.py and render_dashboard_email.py) which are included in the skill package. These scripts are used for data validation and rendering reports into HTML.
[INDIRECT_PROMPT_INJECTION]: The skill processes external task data and notes from Google Sheets. It mitigates injection risks by implementing HTML escaping via a dedicated renderer script before generating email drafts.
Ingestion points: Reads project data from Google Sheets tabs named '專案管理總表' and 'Change & Decision Log'.
Boundary markers: The skill uses a knowledge pack integrity check (SHA256) and schema version detection to ensure data boundaries are respected.
Capability inventory: Uses the google-workspace toolset to read sheets and create Gmail drafts.
Sanitization: Employs html.escape within the render_dashboard_email.py script to sanitize all text before email rendering.
[EXTERNAL_DOWNLOADS]: The documentation includes an installation example referencing a GitHub repository under the author's organization (peter-tu-zynkr), which is a vendor-owned resource.
[DATA_EXPOSURE_AND_EXFILTRATION]: Configuration and project identifiers are resolved from a local file (~/.config/zynkr/pm.json). This follows best practices for secret management by keeping sensitive IDs out of the skill's code and instructions.

project-status-update