The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands (cat and curl) at the end of its workflow to read a local version file and check for updates from the vendor's server.
[EXTERNAL_DOWNLOADS]: Initiates a network request to nextmovetheory.com to verify the current version of the skill. This is a standard vendor-provided resource for maintenance.
[PROMPT_INJECTION]: Uses prescriptive instructions such as "Producer contract (binding)" and employs adversarial reviewer personas for internal self-critique. These are task-specific constraints designed for output quality rather than safety bypasses.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits vulnerability to indirect injection as it ingests untrusted data from user-provided files (e.g., Notion exports, strategy docs) in stage S0.
Ingestion points: Stage S0 (Intake) reads local files and external market research results.
Boundary markers: The skill uses an "Input-as-hypothesis gate" and explicitly tags user-provided data as hypotheses rather than facts.
Capability inventory: The skill has access to subagent creation and shell command execution (curl).
Sanitization: No explicit sanitization or delimiter wrapping for external content is specified beyond the logical treatment of input as "hypotheses."

nmt-craft-value-proposition