The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes content from documentation files (such as design.md and spec.md) and interpolates it into other specifications and index files. This creates a surface for indirect prompt injection where malicious text in the source files could influence the agent's behavior during the archiving process.
Ingestion points: Data is read from openspec/changes/<slug>/design.md and openspec/specs/<capability>/spec.md (specifically Purpose headings).
Capability inventory: The skill performs file system writes and invokes internal tools like /opsx:archive.
Sanitization: No explicit instructions for sanitizing or escaping the content read from these files were found.
Boundary markers: The skill does not utilize specific delimiters to isolate untrusted project data during interpolation.
[COMMAND_EXECUTION]: The skill directs the agent to use command-line utilities for precise file modifications, such as locating specific rows in a markdown table. This introduces a potential surface for command injection if capability names are not strictly validated or escaped before being passed to the shell.
Evidence: Implementation step 25 instructs the agent to "grep for the line whose first cell exactly matches the capability name" when patching the INDEX.md file.

accelint-qrspi-archive