The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of shell-based CLI tools (obsidian-wiki and qmd) to manage its knowledge base. It executes commands such as obsidian-wiki memory profile set, obsidian-wiki memory todo add, and qmd update based on instructions in the skill definition.
[DYNAMIC_EXECUTION]: In Step 5b, the skill generates shell commands by interpolating variables extracted from the conversation into command strings (e.g., obsidian-wiki memory todo add "<the open thread>"). This pattern creates a direct path for conversation data to reach the system shell.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to command injection through the processing of untrusted conversation data. If a user provides input containing shell metacharacters (such as ;, `, $(), or unescaped quotes), and the agent extracts this text for a memory update or a todo item, the resulting shell command will execute the injected payload.
Ingestion points: The skill scans the entire "current conversation or finding" for extraction (SKILL.md Step 1).
Boundary markers: None identified. There are no instructions for the agent to use delimiters or ignore instructions embedded within the conversation text during the extraction phase.
Capability inventory: The agent has the capability to execute shell commands via os.system or equivalent subprocess calls to interact with the obsidian-wiki and qmd tools.
Sanitization: Absent. There are no instructions to escape, validate, or filter the extracted text for shell metacharacters before it is interpolated into the CLI command strings.

wiki-capture