The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill uses dynamic context injection (!command) in SKILL.md to automatically detect the development environment's state when the skill is loaded. These commands (npm list, pip show, and node --version) are benign and used to provide the agent with relevant context about installed Sentry SDK versions.
[EXTERNAL_DOWNLOADS]: The skill provides instructions to update official Sentry SDK packages and bundler plugins from the npm registry and the Python Package Index (PyPI). These downloads target the official @sentry organization and sentry-sdk package, which are well-known services.
[REMOTE_CODE_EXECUTION]: The skill utilizes the official @sentry/migr8 codemod tool via npx to automate breaking change migrations. This involves the execution of code from the official Sentry organization's public repository.
[INDIRECT_PROMPT_INJECTION]: The skill has an inherent attack surface for indirect prompt injection because its core functionality requires reading and processing untrusted source code from the project directory (src/) to identify deprecated API usage.
Ingestion points: The agent scans project source files (.ts, .js, .py) using grep commands to find specific patterns.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to prevent it from following commands that might be embedded within the source code files it reads.
Capability inventory: The skill is granted permissions to edit files (Write, Edit) and execute shell commands via Bash (npm, npx, pip, node).
Sanitization: The skill does not implement validation or sanitization of the content extracted from the source files before the agent processes it.

sentry-upgrade-migration