The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions in SKILL.md repeatedly direct the agent to execute shell commands using the story CLI tool. These instructions interpolate user-provided values directly into shell strings.
Evidence in SKILL.md (Creating a Location): story names "{Candidate}" and story add location "{Location Name}" --type "{type}".
Evidence in SKILL.md (Creating A Faction): story add faction "{Faction Name}" --type "{...}".
Evidence in SKILL.md (Creating An Artifact): story add artifact "{Artifact Name}" --type "{...}".
This pattern allows for command injection if a user provides a name containing shell metacharacters (e.g., "; rm -rf /; ").
[DYNAMIC_EXECUTION]: The skill provides a fallback mechanism for execution when the primary CLI is not installed, instructing the agent to run a JavaScript file from a relative path outside the skill's own directory.
Evidence in SKILL.md: "use the bundled fallback node ../story-maintenance/scripts/story.js with the same arguments".
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external markdown files which could contain untrusted data that influences the agent's behavior.
Ingestion points: The agent reads story.md, worldbuilding/_index.md, and various element files in the worldbuilding/ directory.
Boundary markers: There are no instructions for the agent to use delimiters or ignore instructions embedded within the ingested worldbuilding files.
Capability inventory: The skill has significant capabilities including file system writes and shell command execution via the story CLI.
Sanitization: The skill lacks instructions for sanitizing or escaping content read from files or provided by the user before it is used in logic or shell commands.

worldbuilding