The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external event debrief logs, which creates a potential surface for indirect prompt injection if the logs contain adversarial instructions.
Ingestion points: Data enters the agent context via dated debrief logs provided by the user or a sibling skill, as described in the 'Interview' and 'Workflow' sections of SKILL.md.
Boundary markers: The skill implements a 'signal gate' (defined in references/filing-gate-and-trend-evidence.md) that requires findings to be current, reproducible, and generalizing, which serves as a logical filter for incoming data.
Capability inventory: The skill instructions do not contain any subprocess calls, dynamic code execution, network operations, or file-writing capabilities. Its primary function is text analysis and reporting within the agent's context.
Sanitization: The skill mandates 'de-personalisation' in references/filing-gate-and-trend-evidence.md, requiring the agent to rewrite entries to describe conditions rather than people, which acts as a form of content normalization.
[REMOTE_CODE_EXECUTION]: No remote code execution patterns, package installations, or script downloads were detected. References to samber/dev-event-organizer-skills are internal pointers to sibling skills by the same author.
[DATA_EXFILTRATION]: No network operations or sensitive file access patterns were identified. The skill's focus is on organizational memory and trend detection within the user's provided data.

event-continuous-improvement