The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill acts as a surface for indirect prompt injection by processing external data from WordPress audit logs without explicit sanitization steps.
Ingestion points: The skill ingests data from the respira_generate_activity_report tool, specifically processing a notes array containing arbitrary strings describing site changes.
Boundary markers: There are no explicit delimiters (e.g., XML tags) or instructions provided to the agent to prevent the interpretation of instructions contained within the audit log notes.
Capability inventory: The skill possesses the ability to write to the site database using respira_create_custom_post and perform network requests to the author's API.
Sanitization: No sanitization or validation of the audit log content is mentioned before it is interpolated into the report generation prompts.
[DATA_EXFILTRATION]: The skill transmits site metadata and activity statistics to the author's server.
Evidence: A POST request is sent to https://www.respira.press/api/skills/track-usage containing site URL hashes, activity totals, and framing choices.
The telemetry implementation follows safe practices by explicitly excluding sensitive information such as client names, specific notes, or report bodies.

activity-report-composer