All skills
dianel555 avatar

/codex-cc

@b6a6d5f
by dianeldianel555/dskills65 stars
8

Delegates coding tasks from Codex to local Claude Code in print mode while preserving Claude's normal runtime customizations by default. Use when: (1) You want Codex to call Claude Code locally, (2) You need Claude-side skills, plugins, MCP servers, custom commands, CLAUDE.md rules, or workspace settings to stay active, (3) You want resumable Claude sessions from Codex via SESSION_ID, (4) You need thin `claude mcp` / `claude plugin` passthrough from the same bridge.

Use this Skill: https://skilld.dev/gh/dianel555/dskills/codex-cc

This session only. Nothing lands on disk.

SKILL.md

≈120 tokens always: the name and description. ≈788 when used: this file. ≈1.2k more on demand in 2 files.

Quick Start

python scripts/claude_bridge.py --cd "/path/to/project" --PROMPT "Analyze auth flow"

Output: JSON with success, SESSION_ID, agent_messages, stream_file (the raw Claude JSONL stream), and optional all_messages, stderr, or error.

Runtime Contract

  • Default execution uses claude -p and does not force --safe-mode, --bare, --disable-slash-commands, or --strict-mcp-config.
  • Claude Code therefore keeps its normal loading path for trusted-workspace customizations such as CLAUDE.md, skills, plugins, MCP servers, custom commands, and rules.
  • claude -p skips the interactive trust dialog and silently ignores invalid settings files, so use this only in workspaces you already trust and whose .claude settings already validate.
  • The bridge requests stream-json, persists each record immediately, and only treats Claude's final result record as a completed answer. Intermediate assistant records never mask a failed or incomplete turn.
  • On Windows the PROMPT is delivered through stdin when claude resolves to a .cmd/.bat shim, so the prompt never passes through cmd.exe quoting and cannot hit the command-line length limit.

Parameters

usage: claude_bridge.py [-h] --PROMPT PROMPT --cd CD [--SESSION_ID SESSION_ID]
                        [--model MODEL]
                        [--permission-mode {,acceptEdits,auto,bypassPermissions,manual,dontAsk,plan}]
                        [--dangerously-skip-permissions] [--timeout TIMEOUT]
                        [--stream-file STREAM_FILE] [--return-all-messages]
                        {mcp,plugin} ...

options:
  --PROMPT PROMPT                Instruction for the task to send to Claude Code.
  --cd CD                        Workspace root for Claude Code (cwd + --add-dir).
  --SESSION_ID SESSION_ID        Resume a conversation by session UUID.
  --model MODEL                  Claude model override. Aliases resolve against the active endpoint.
  --permission-mode ...          Claude permission mode override.
  --dangerously-skip-permissions Bypass Claude permission checks.
  --timeout TIMEOUT              Bridge-level timeout in seconds. Omit for no bridge deadline.
  --stream-file STREAM_FILE      Raw Claude stream-json JSONL destination.
  --return-all-messages          Include parsed stream records in the result.

subcommands:
  mcp                            Thin passthrough to `claude mcp`.
  plugin                         Thin passthrough to `claude plugin`.

Sessions

Capture SESSION_ID from the first successful response and reuse it for follow-ups. SESSION_ID is empty when a run never established a session (claude produced no stream output): resume only with an id the bridge actually returned.

# New Claude session
python scripts/claude_bridge.py --cd "/project" --PROMPT "Inspect failing tests"

# Resume the same Claude session
python scripts/claude_bridge.py --cd "/project" --SESSION_ID "uuid-from-response" --PROMPT "Now propose the fix"

Passthrough

python scripts/claude_bridge.py mcp list
python scripts/claude_bridge.py plugin list

Source: SKILL.md on GitHub

1 warning18d3 checks · Risk MEDIUM
  • Gen Agent Trust Hub18d

    This skill acts as a bridge to execute the local Claude Code CLI. It includes a specific flag that bypasses all security prompts and permission checks, which could allow an agent to perform unauthorized actions on the local system if misused.

  • Socket18d

    No alerts

  • Snyk18d

    Risk: LOW · No issues

Signed by skilld at b6a6d5f. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub last week.

Activeupdated 2 weeks ago

README badge

README badge for dianel555/dskills/codex-cc