The Agent Skills Directory

[COMMAND_EXECUTION] (LOW): The skill executes a local shell command to retrieve dependency information.
Evidence: !grep -A 100 '^\\[dependencies\\]' Cargo.toml 2>/dev/null | head -30 || echo "No Cargo.toml found" in SKILL.md.
Analysis: While this command is used to provide context to the agent, it represents an active shell invocation during skill processing.
[PROMPT_INJECTION] (LOW): Indirect prompt injection vulnerability through external data ingestion.
Ingestion points: Reads Cargo.toml via shell command.
Boundary markers: Absent; the file content is injected directly into the prompt context.
Capability inventory: The skill provides architectural advice but the environment processing the skill allows shell command execution.
Sanitization: Absent; the content of Cargo.toml is not escaped or validated before ingestion. An attacker with write access to the local project's Cargo.toml could attempt to influence the agent's behavior.

m11-ecosystem