The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on shell commands to interact with the openspec CLI (e.g., openspec new change, openspec status, openspec instructions). These commands are used to scaffold project directories, monitor artifact status, and retrieve JSON-formatted instructions for file generation.
[INDIRECT_PROMPT_INJECTION]: The skill processes instructions and data retrieved from external sources, which could potentially contain malicious content designed to influence the agent's behavior.
Ingestion points: The skill ingests data from two primary external sources: the JSON output of the openspec instructions command (which includes context, rules, and instruction fields) and the content of completed dependency artifacts read from the filesystem.
Boundary markers: The instructions do not define boundary markers or explicit delimiters to separate the external content from the agent's system prompt or to warn the agent to ignore any embedded instructions within that content.
Capability inventory: The skill has the capability to execute shell commands via the openspec CLI and write arbitrary content to the local filesystem at paths provided by the CLI tool (resolvedOutputPath).
Sanitization: No evidence of sanitization, filtering, or validation is present for the data retrieved from the CLI or the dependency files before it is used to guide the generation of new files.

openspec-propose