The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands to determine repository state and configuration.
Runs git remote get-url origin to resolve the source repository URL.
Runs git rev-parse --abbrev-ref HEAD to identify the current working branch.
[INDIRECT_PROMPT_INJECTION]: The skill ingests markdown content from the local workspace to generate structured documentation for language models, which could include malicious instructions.
Ingestion points: Reads all markdown files within the wiki directory and root directory to populate llms.txt and llms-full.txt (SKILL.md).
Boundary markers: Uses <doc> XML-style tags with title and path attributes to delimit content in the generated llms-full.txt file.
Capability inventory: The skill has the ability to execute shell commands (git), read workspace files, and write new files to the filesystem (llms.txt, wiki/llms.txt, wiki/llms-full.txt).
Sanitization: The instructions explicitly mention stripping YAML frontmatter (--- blocks) but otherwise preserve the full content of the markdown files, including diagrams and citations, without further filtering or validation of the embedded text.

wiki-llms-text