The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill recommends installing a third-party NPM package @colbymchenry/codegraph to enable enhanced codebase mapping features. It correctly includes instructions for the AI agent to seek explicit user permission before initiating the installation.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted codebase data to generate its focus maps, creating a potential surface for indirect prompt injection if source files contain malicious instructions.
Ingestion points: The obsidian-wiki code-understand tool reads and parses source code files and directory structures from the local filesystem.
Boundary markers: No explicit delimiters or instructions are provided to the agent to help distinguish between codebase content and system-level directives in the tool's output.
Capability inventory: The skill has the ability to read arbitrary project files, perform regex-based extraction, execute the obsidian-wiki CLI, and suggest shell commands like npm install.
Sanitization: No evidence of input filtering or content sanitization is present in the skill's instructions or primary logic.
[COMMAND_EXECUTION]: Utilizes the obsidian-wiki CLI and standard package management tools like npm to perform its core functions and manage dependencies.

code-understand