The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash(openspec:*) tool to execute various commands via the OpenSpec CLI.
Evidence includes calls to openspec list, openspec status, openspec new change, and openspec instructions.
It utilizes dynamic arguments such as --store <id> and --change <name> derived from project state and user input.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to read and process external data from the codebase, creating a risk surface where malicious instructions in code or configuration could influence agent behavior.
Ingestion points: Reads project configuration from openspec/config.yaml (or .yml), explores codebase source files, and parses JSON output from the openspec CLI.
Boundary markers: The instructions explicitly tell the agent not to copy content from the context files into artifacts and to treat them as constraints, but no formal delimiters or "ignore instructions" tags are used for file ingestion.
Capability inventory: The skill has the ability to execute shell commands (via the openspec CLI) and write new markdown artifacts to the local file system.
Sanitization: There is no mention of sanitization, validation, or filtering of the content read from the codebase before it is processed by the agent.

openspec-explore