The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses npx @alwaysmeticulous/cli to identify and download session recordings. This is the primary mechanism for interacting with the Meticulous platform.
[EXTERNAL_DOWNLOADS]: Fetches structured session data, including user interaction logs and network mocks, from the Meticulous service to the local .meticulous/sessions/ directory.
[CREDENTIALS_UNSAFE]: The skill downloads and instructs the agent to read files containing browser state information, specifically storage/cookies.json, storage/local-storage.json, and storage/session-storage.json. These files may contain sensitive session tokens or authentication data from the recorded sessions.
[PROMPT_INJECTION]: Ingesting and analyzing external session data (user events, network requests, and storage) creates a surface for Indirect Prompt Injection. If a recorded session includes attacker-controlled content, it could contain instructions designed to manipulate the agent.
Ingestion points: Multiple JSON files in .meticulous/sessions/, including user-events.json and network-requests/summary.json.
Boundary markers: None identified. The agent is not instructed to use specific delimiters or to disregard potential instructions within the session data.
Capability inventory: The skill allows for command execution (npx) and file system navigation.
Sanitization: No explicit sanitization or validation of the downloaded JSON content is performed before the agent reads it.

use-session-data