operon-fs-execs
Use when an agent needs to read, write, copy, remove, or mount remote files, or run and interact with execs on Operon nodes.
Works with
Agent Skills format with YAML frontmatter. Claude Code reads it as-is.
--- name: "operon-fs-execs" description: "Use when an agent needs to read, write, copy, remove, or mount remote files, or run and interact with execs on Operon nodes." license: "MPL-2.0" --- # Operon FS And Execs Use this skill for filesystem and process execution workflows. Graph steps use `exec.run` for non-interactive execution. Filesystem targets use `node:/path` form. Always inspect the exact command help before running a command: ```bash operon fs --help operon fs read --help operon fs write --help operon exec run --help operon exec session --help operon exec logs --help operon exec stdin --help ``` Filesystem command choice: - Inspect metadata: `operon fs stat <node:/path>`. - List a directory: `operon fs list <node:/path>`. - Stream a remote file: `operon fs read <node:/path>`. - Write inline content or a local file: `operon fs write <node:/path>`. - Create a directory: `operon fs mkdir <node:/path>`. - Copy: `operon fs copy <node:/from> <node:/to>`. - Rename or move: `operon fs rename <node:/from> <node:/to>`. - Truncate: `operon fs truncate <node:/path> --size <bytes>`. - Remove: `operon fs rm <node:/path>`. - Linux local mount: `operon mount <node:/path> --to <mountpoint>`. Confirm before `write`, `copy` over an existing destination, `rename`, `truncate`, `rm`, or mounting over a non-empty directory. Then verify with `fs stat`, `fs list`, and `audit show`. Exec command choice: - Run a command and wait: `operon exec run <node> -- <command>`. - Start and return immediately: `operon exec run <node> --detach -- <command>`. - Open a PTY-backed interactive command when policy allows sessions: `operon exec session <node> -- <command>`. The CLI uses the attached terminal size by default and forwards Unix resize events during interactive sessions. - Check status: `operon exec status <node> <exec_id>`. - Read or follow stdout/stderr: `operon exec logs <node> <exec_id>`. - Send stdin or close it: `operon exec stdin <node> <exec_id>`. - Cancel: `operon exec cancel <node> <exec_id>`. Use `exec.run` for non-interactive commands and graph steps. Use `exec.session` only when the command needs terminal semantics such as shell line editing, terminal control, or PTY-aware programs. Treat exec execution as policy-sensitive. Confirm cwd, timeout, secrets, and destructive shell commands. Check `audit show` after running execs that mutate files, services, or external systems.
More General & Other skills
find-skills
vercel-labs/skills
Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.
grill-me
mattpocock/skills
A relentless interview to sharpen a plan or design.
grill-with-docs
mattpocock/skills
A relentless interview to sharpen a plan or design, which also creates docs (ADR's and glossary) as we go.

