Agents burn 3-5 LLM calls per REST interaction fumbling with curl syntax, JSON escaping, and auth headers. This pattern wraps your app's API in a domain-specific CLI that reads Eve's auto-injected environment variables and turns `curl -X POST -H "Authorization: Bearer $TOKEN" ...` into `eden changeset create --file data.json`. You bundle it with esbuild into a single executable, commit it to your repo, and declare it in your manifest. The platform puts it on PATH for any agent with `with_apis` access. The auto-detection pattern is clever: if there's only one project, commands skip the `--project` flag and just work. It's the difference between agents that guess at API calls and agents that run actual commands.
npx -y skills add incept5/eve-skillpacks --skill eve-app-cli --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
sentry/dev
degausai/wonda
github/awesome-copilot
insforge/agent-skills
github/awesome-copilot
breaking-brake/cc-wf-studio