codex-ralph 0.6.4 → 0.7.0

package/README.md

@@ -31,6 +31,7 @@ Optional flags:
 --max-iterations=10
 --cursor-agent
 --gemini-agent
+--claude-agent
 ```
 
 Message format (session emoji is chosen randomly from a small fixed set and stays constant for the session):
@@ -50,6 +51,12 @@ Each requirement is a level-2 heading with a checkbox. The loop picks the first
 
 Description: Free-form task details.
 
+Acceptance criteria: Free-form acceptance criteria.
+
+## [ ] Requirement description
+
+Description: Free-form task details.
+
 Acceptance criteria: Free-form acceptance criteria.
 ```
 
@@ -64,23 +71,3 @@ Acceptance criteria: Free-form acceptance criteria.
 - Marks items complete only when all steps are satisfied.
 - Uses conventional commits for each completed requirement (no story ID requirement).
 - If no local `AGENTS.md` exists up the tree, use `./.codex/AGENTS.md`. In monorepos, prefer package-level AGENTS.md for package-specific knowledge.
-
-## Changelog
-
-### 0.6.4
-- Fix: Removed conflicting `--yolo` flag from Gemini agent command.
-
-### 0.6.3
-- Feat: Added custom flags (--yolo, --no-sandbox, --approval-mode yolo) for Gemini agent command.
-
-### 0.6.2
-- Fix: Further corrected Gemini agent command to use positional argument instead of stdin.
-
-### 0.6.1
-- Fix: Corrected the command for the Gemini agent.
-
-### 0.6.0
-- Add support for Gemini agent via `--gemini-agent` flag.
-
-### 0.5.0
-- Version bump.

package/agent/prompt.md

@@ -49,7 +49,7 @@ Only add patterns that are **general and reusable**, not story-specific details.
 Before committing, check if any edited files have learnings worth preserving in nearby AGENTS.md files:
 
 1. **Identify directories with edited files** - Look at which directories you modified
-2. **Check for existing AGENTS.md** - Look for AGENTS.md in those directories or parent directories. If none is found all the way up, use `./.codex/AGENTS.md`.
+2. **Check for existing AGENTS.md/CLAUDE.md** - Look for AGENTS.md/CLAUDE.md in those directories or parent directories
 3. **Monorepos** - If working under `./packages/<name>` (or similar), prefer placing AGENTS.md in that package if the knowledge is package-specific.
 4. **Add valuable learnings** - If you discovered something future developers/agents should know:
   - API patterns or conventions specific to that module

package/agent/ralph-loop.sh

@@ -14,6 +14,7 @@ NOTES_FILE=""
 SESSION_EMOJI=""
 USE_CURSOR_AGENT=false
 USE_GEMINI_AGENT=false
+USE_CLAUDE_AGENT=false
 
 session_emoji() {
   if [[ -n "${SESSION_EMOJI}" ]]; then
@@ -45,13 +46,14 @@ send_telegram() {
 
 usage() {
   cat <<USAGE
-Usage: ralph-loop SPRINT_PATH [--max-iterations=N] [--cursor-agent]
+Usage: ralph-loop SPRINT_PATH [--max-iterations=N] [--cursor-agent] [--gemini-agent] [--claude-agent]
 
 Options:
   SPRINT_PATH            Path to the sprint markdown file (required).
   --max-iterations=N     Stop after N iterations (0 = no limit, default 10).
   --cursor-agent         Use cursor-agent instead of codex (default: disabled).
   --gemini-agent         Use gemini-agent instead of codex (default: disabled).
+  --claude-agent         Use claude instead of codex (default: disabled).
   -h, --help             Show this help.
 USAGE
 }
@@ -69,6 +71,9 @@ for arg in "$@"; do
     --gemini-agent)
       USE_GEMINI_AGENT=true
       ;;
+    --claude-agent)
+      USE_CLAUDE_AGENT=true
+      ;;
     -h|--help)
       usage
       exit 0
@@ -257,6 +262,8 @@ while true; do
     cursor-agent -f -p < "$prompt_tmp"
   elif [[ "$USE_GEMINI_AGENT" == "true" ]]; then
     gemini --no-sandbox --approval-mode yolo "$(cat "$prompt_tmp")"
+  elif [[ "$USE_CLAUDE_AGENT" == "true" ]]; then
+    claude -p --dangerously-skip-permissions < "$prompt_tmp"
   else
     codex exec - < "$prompt_tmp"
   fi

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "codex-ralph",
-  "version": "0.6.4",
+  "version": "0.7.0",
   "description": "Ralph Loop sprint runner for Codex.",
   "bin": {
     "ralph-loop": "agent/ralph-loop.sh"