Read and execute task delegated from Claude. Use when user says read task, get task, or pick up claude task.
Read and execute a task delegated from Claude Code.
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.
Write working to $AGENT_COLLAB_DIR/status
Read $AGENT_COLLAB_DIR/requests/task.md and parse:
Also read $AGENT_COLLAB_DIR/context/shared.md for project context.
Perform thorough code review:
Be thorough - take your time.
Implement the requested feature:
Take time for high-quality implementation.
Critically analyze the plan:
Write complete response to $AGENT_COLLAB_DIR/responses/response.md:
# Codex Response
## Task Type
[Task type handled]
## Completed At
[Timestamp]
## Summary
[Brief summary]
## Detailed Findings/Output
[For CODE_REVIEW:]
### Critical Issues
[List with severity]
### Security Concerns
[Security issues]
### Suggestions
[Improvements with examples]
---
[For IMPLEMENT:]
### Implementation Overview
[What was implemented]
### Files Created/Modified
[List of files]
### Integration Notes
[How to integrate]
---
[For PLAN_REVIEW:]
### Overall Assessment
[Is plan sound?]
### Concerns
[Issues identified]
### Recommendations
[Specific suggestions]
Write done to $AGENT_COLLAB_DIR/status
Tell user response is ready and Claude can read it with /codex-read.