The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection by reading untrusted project configuration files to extract verification commands.
Ingestion points: The agent reads package.json, Cargo.toml, pyproject.toml, go.mod, and Makefile from the project root.
Boundary markers: No explicit delimiters or boundary markers are used to isolate the data read from these files.
Capability inventory: The skill has the capability to write files (openspec/VERIFY.md, openspec/GLOSSARY.md), execute the openspec CLI, and install global NPM packages.
Sanitization: No automated sanitization is performed on the extracted text, although the instructions mandate user confirmation before writing the proposed commands to a file.
[EXTERNAL_DOWNLOADS]: The skill downloads and installs the OpenSpec CLI from the public NPM registry.
Fetches the @fission-ai/openspec package via npm install -g.
The installation is guarded by an explicit requirement for user confirmation.
[COMMAND_EXECUTION]: The skill executes local shell commands to check the environment and initialize the workspace.
Executes openspec --version to verify the local installation.
Executes openspec init --no-animation --tools none to create the project directory structure.

openspec-setup