The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for the agent to recommend a shell command to the user for installing complementary skills. Evidence: npx @nextstage-brasil/harness --preset complements --yes found in SKILL.md and references/skill-integrations.md. This command targets a package provided by the vendor ('nextstage-brasil').
[EXTERNAL_DOWNLOADS]: The skill references the installation of an external package from the vendor's registry. Evidence: @nextstage-brasil/harness in SKILL.md. This is a legitimate vendor resource associated with the skill's author.
[PROMPT_INJECTION]: The skill processes data from various local project files which could potentially contain untrusted instructions. This represents an indirect prompt injection attack surface.
Ingestion points: Files in docs/versions/ (e.g., requirements.md, tasks/*.md) and docs/context/ (e.g., design-brief.md, architecture-rules.md).
Boundary markers: No specific delimiters or markers to isolate external content are defined in the instructions.
Capability inventory: The skill can delegate execution to other worker skills and suggest the execution of shell commands.
Sanitization: No explicit sanitization or validation of the ingested content is mentioned.
[NO_CODE]: The skill package contains documentation, configuration, and evaluation files, but does not include any executable scripts or binaries within its own folder structure.

nextstage-sdd