antorsae avatar

codex-read

Read and integrate response from Codex. Use when user says codex read, get codex response, or check

by antorsae|Open Source

Codex Read Skill

Read the response from Codex and integrate findings.

Steps

Before any file operations, resolve the .agent-collab directory so commands work outside the project root:

AGENT_COLLAB_DIR="${AGENT_COLLAB_DIR:-}"
if [ -n "$AGENT_COLLAB_DIR" ]; then
  if [ -d "$AGENT_COLLAB_DIR/.agent-collab" ]; then
    AGENT_COLLAB_DIR="$AGENT_COLLAB_DIR/.agent-collab"
  elif [ ! -d "$AGENT_COLLAB_DIR" ]; then
    AGENT_COLLAB_DIR=""
  fi
fi

if [ -z "$AGENT_COLLAB_DIR" ]; then
  AGENT_COLLAB_DIR="$(pwd)"
  while [ "$AGENT_COLLAB_DIR" != "/" ] && [ ! -d "$AGENT_COLLAB_DIR/.agent-collab" ]; do
    AGENT_COLLAB_DIR="$(dirname "$AGENT_COLLAB_DIR")"
  done
  AGENT_COLLAB_DIR="$AGENT_COLLAB_DIR/.agent-collab"
fi

If $AGENT_COLLAB_DIR does not exist, stop and ask for the project root.

1. Check Status

Read $AGENT_COLLAB_DIR/status:

  • idle: No pending task. Inform user.
  • pending: Codex hasn't started. Check Codex pane.
  • working: Codex still processing. Wait.
  • done: Proceed to read response.

2. Read Response

Read $AGENT_COLLAB_DIR/responses/response.md and parse:

  • Task type completed
  • Findings/output
  • Any code produced
  • Recommendations

3. Present to User

Summarize clearly based on task type:

For CODE_REVIEW:

  • Bugs/issues with severity
  • Security concerns
  • Performance issues
  • Suggested improvements

For IMPLEMENT:

  • What was implemented
  • Approach taken
  • Integration instructions

For PLAN_REVIEW:

  • Critique summary
  • Identified risks
  • Alternative suggestions

4. Integrate (if applicable)

For implementations:

  • Ask if user wants to integrate
  • Apply changes if yes
  • Run relevant tests

5. Reset Status

Write idle to $AGENT_COLLAB_DIR/status

6. Next Steps

Suggest appropriate follow-up:

  • Reviews: Address issues found
  • Implementations: Test and validate
  • Plan reviews: Iterate on plan
codex-read - AI Agent Skill for Claude Code & Cursor | Agent Skills