The Agent Skills Directory

[SAFE]: The skill defines a structured workflow for data aggregation and reporting. It follows a predictable template and does not perform any sensitive system operations.
[COMMAND_EXECUTION]: While the skill mentions calling other skills like focus-keeper and pipeline-manager, these are treated as internal agent functions and do not involve raw shell command execution or subprocess spawning.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from potentially untrusted sources such as outreach-feedback.jsonl (external responses) and pr-opportunities.json (found PR opportunities).
Ingestion points: memory/outreach-feedback.jsonl, memory/pr-opportunities.json, and memory/github-stars.json.
Boundary markers: Absent. The skill does not explicitly instruct the agent to ignore instructions embedded within these data sources.
Capability inventory: The skill is limited to reading files and generating a text report. It performs logging to autonomous-actions.jsonl, but has no network, file-deletion, or code execution capabilities.
Sanitization: Absent. Content is interpolated directly into the brief.
Context: Because the output is a human-readable brief for a CEO and the skill lacks dangerous side-effect capabilities, this attack surface is considered low-risk and does not escalate the verdict.

morning-brief