The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill instructions direct the agent to read from sensitive files, specifically ~/.obsidian-wiki/config and .env. While this is intended to locate the user's Obsidian vault path, these files are standard locations for storing API keys, tokens, and other sensitive credentials.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to its core function of processing untrusted conversation data and writing it to the local file system.
Ingestion points: The "current conversation" history is ingested to identify knowledge worth preserving.
Boundary markers: Absent. The skill does not provide instructions to the agent to distinguish between user-provided data and system-level instructions or to ignore embedded commands.
Capability inventory: The skill possesses capabilities to read the entire wiki vault and configuration files, create new markdown notes, and modify existing tracking files like index.md, log.md, and hot.md.
Sanitization: Absent. There is no specified logic for validating or escaping the content extracted from the conversation before it is written to permanent files.

wiki-capture