The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use shell commands such as pnpm evlog, npx evlog, and npm exec to query logs, check system status, and initialize the environment.
[EXTERNAL_DOWNLOADS]: The skill suggests using npx, bunx, or npm exec to fetch and execute the @evlog/cli package and related dependencies from external registries if they are not already present in the project.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze log files (.jsonl) generated by the application, which represents a surface for indirect prompt injection.
Ingestion points: The agent reads structured NDJSON/JSONL events from the .evlog/logs/ directory, including application-specific context and error messages.
Boundary markers: No specific boundary markers or instructions to ignore embedded commands within the log data are provided.
Capability inventory: The agent has the ability to execute shell commands (pnpm, npx, npm), read local files, and potentially make network requests via the evlog CLI or memory drain endpoints.
Sanitization: There is no evidence of sanitization or filtering of the log content before the agent is tasked with summarizing and explaining the events.
[DATA_EXFILTRATION]: The skill directs the agent to access potentially sensitive information stored in logs, including stack traces, request paths, and custom business context such as user information and payment details, which are then processed and summarized for the user.

analyze-logs