The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest external instructions and treat them as behavior contracts. If these instructions contain malicious patterns, they could influence agent behavior during the workflow.
Ingestion points: The skill reads JSON output from openspec instructions archive and openspec instructions specs, specifically extracting context and operationGuidance fields.
Boundary markers: While it provides logic to handle conflicts with built-in steps, it lacks strict delimiters or instructions to ignore potential adversarial commands embedded in the guidance text.
Capability inventory: The skill possesses the capability to execute shell commands (mkdir, mv) and interact with the openspec CLI, which are guided by the ingested text.
Sanitization: There is no evidence of sanitization, filtering, or validation of the natural language instructions received from the CLI outputs before they are processed by the agent.
[COMMAND_EXECUTION]: The skill uses the Bash tool to perform file system operations (mkdir -p, mv) and to interact with the openspec CLI for managing workflow states. These actions are within the expected scope of an archival skill.

openspec-archive-change