The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes git remote get-url origin to identify the repository source URL, which is then included in the exported data.
[DATA_EXFILTRATION]: The skill aggregates summarized session content ('learnings') along with the local repository URL for transmission to an external 'Post' mechanism. The destination of this data is not explicitly defined within the skill instructions.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it scans and processes the entire session history—which may contain untrusted data from external sources—to generate its output.
Ingestion points: The process 'Scan the session' ingests all previous interactions and content within the active agent session (SKILL.md).
Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands within the harvested session data.
Capability inventory: The skill utilizes shell command execution via git and performs network-based data export ('Post').
Sanitization: The instructions include a manual heuristic for the AI to 'Drop' secrets and PII, but no programmatic validation or sanitization is present.

reflect