The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill scans local project files to identify Sentry initialization code, creating a surface for indirect prompt injection from codebase content.
Ingestion points: grep command in SKILL.md extracts code snippets from .ts, .js, .mjs, and .py files.
Boundary markers: No explicit delimiters or instructions are used to separate scanned code from the agent's primary instructions.
Capability inventory: The skill has access to tools including npm, node, curl, npx, sentry-cli, and python3 via shell execution.
Sanitization: No escaping or validation is performed on the extracted file content before it is processed by the agent.
[REMOTE_CODE_EXECUTION]: The skill fetches health data from Sentry's official status API and processes the response using a piped Python script for diagnostic purposes.
[DYNAMIC_CONTEXT_INJECTION]: Benign environment discovery commands are executed at skill load time using the ! syntax to detect the presence of Node.js and Python SDKs as well as the Sentry CLI.
[DYNAMIC_EXECUTION]: The skill uses node -e and python3 -c to execute small diagnostic script templates for validating environment configuration and event delivery.

sentry-common-errors