skillZs
★ LIVE SKILL TAGS ★
>>> LIVE SKILLS INDEX <<<
* OPEN SOURCE *
NO LOGIN, NO TRACKING
※ REAL INSTALL DATA ※
← back to all skills
robsonrung/rar-skills76 installs

codex-runner

Execute an external Codex CLI prompt in non-interactive exec mode. Use only for an explicit Codex CLI request, a selected Astra, Sol, or Terra route unavailable as native host delegation, or an approved external fallback.

How do I install this agent skill?

npx skills add https://github.com/robsonrung/rar-skills --skill codex-runner
view source ↗

Is this agent skill safe to install?

  • Gen Agent Trust Hubpass

    The skill is a robust execution wrapper for the Codex CLI, providing a controlled environment for running prompts with role-based instructions and session persistence. It incorporates significant security mitigations, including a default read-only sandbox and a mechanism to verify and disable external tool plugins (MCP servers) before execution.

  • Socketpass

    No alerts

  • Snykpass

    Risk: LOW · No issues

What does this agent skill do?

Codex Runner

Use shared/references/model-preview.md to choose the branch first. Deterministic checks need no model selection. Actual dispatch uses one concrete approved snapshot; nested calls consume it without re-resolution or extra workers.

Execute the caller's scoped prompt through the local CLI. Shared roles, envelope keys, and result handling live in shared/references/runner-common.md. Preserve seat fidelity: name the provider that actually answered, and report unavailable or unverified seats honestly.

Native routing

Read shared/references/host-model-execution.md before choosing an external route. If the current host can dispatch the exact selected Astra, Sol, or Terra model with the required effort, isolation, and receipt policy in a native subagent or task thread, use that route. agents/openai.yaml is host UI metadata; it does not dispatch a native model.

Invocation

Set SKILL_DIR to this loaded skill directory in the same shell call:

SKILL_DIR="<absolute path of this skill directory>";
python3 "$SKILL_DIR/scripts/run_codex.py" "<scoped prompt>" \
  --model <approved-model-id> --effort <approved-effort> --disable-fallback --json

Copy the exact model and effort from the selected snapshot. Add its role, tool mode, receipt policy, and limits. Omit an effort flag only for a selected runtime controlled route. Nested calls use these saved options without resolving defaults again. The example does not grant source sharing or tool authority.

Runtime boundaries

  • Use the local authenticated codex CLI in exec mode. Analysis roles default to a read-only sandbox.
  • --tool-profile repo_read_only also disables plugins, apps, web search, and MCP servers, and records a configured tool receipt. codex exec reports no startup tools, so this receipt attests configuration, not observation. User Codex hooks still run; they are not model tools. Codex cannot run no_tools.
  • --allow-write, an explicit sandbox, or --full-auto can change that boundary only within the caller's authority. Full auto requires approval for an unattended run.
  • Prompt and file context may be sent to the configured provider. Set --working-dir to the task's repository or package.
  • The wrapper forwards model and effort. Current headless output does not prove the serving model; a configured label alone gives an unverified receipt.
  • Direct calls can fall back once to claude-runner when the CLI is missing. Approved routes disable fallback. Native resume cannot fall back to another runner.
  • Read references/prompting.md only when a prompt needs clarification or an output contract needs design. It is optional guidance, not a required recipe.

Load by need

Result

Read agent_message and the wrapper envelope. Check success, serving-model receipt, effective runner/provider, and any fallback or malformed-output status. A completed process alone does not prove the requested task or model was used. Report touched files when the caller authorized edits.

Add the canonical catalog link to the repository README so users can inspect current installs and available audits. The publishing guide covers the complete discovery path.

<a href="https://skillzs.dev/skills/robsonrung/rar-skills/codex-runner">View codex-runner on skillZs</a>