The Agent Skills Directory

[DATA_EXFILTRATION]: The skill instructs the agent to access the .env file, which is categorized as a sensitive file path.
Evidence: Found in SKILL.md in the 'Before You Start' section: 'Read .env to get OBSIDIAN_VAULT_PATH'.
Security Concern: .env files are standard locations for storing sensitive environment variables, API keys, and credentials. Reading this file exposes its entire contents to the agent's context, posing a high risk of accidental secret exposure.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by processing untrusted data from the user's Obsidian vault.
Ingestion points: The skill iterates through all markdown (.md) files in the vault, as well as index.md and log.md (documented in SKILL.md).
Boundary markers: Absent. There are no instructions to use delimiters or to treat the note content as data that should not be interpreted as instructions.
Capability inventory: The skill has significant capabilities, including reading and writing files, creating new pages, modifying frontmatter, and triggering other skills like cross-linker.
Sanitization: Absent. The skill directly compares claims across notes and updates page content based on its analysis of the raw text.

wiki-lint