The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx to execute @claude-flow/cli, which involves downloading and running external code from the NPM registry at runtime.
[COMMAND_EXECUTION]: The skill executes local shell scripts (sparc-init.sh and sparc-review.sh) to perform file system operations such as creating directories and files.
[INDIRECT_PROMPT_INJECTION]: The skill contains potential command injection surfaces where user-supplied data is used in shell execution contexts.
Ingestion points: User input is accepted through placeholders in SKILL.md (e.g., [requirements], [feature], [design]) and as positional arguments passed to the shell scripts.
Boundary markers: There are no boundary markers or instructions to treat user input as non-executable data, allowing potential escape from the intended command string.
Capability inventory: The skill possesses the capability to execute arbitrary shell commands and perform file system manipulations through its scripts and CLI calls.
Sanitization: No input validation, escaping, or sanitization logic is evident in the scripts or the skill instructions before user input is passed to the shell.

sparc-methodology