Claude Code returns "API Error: <model>'s safeguards flagged this message" as a success-typed result. EJClaw delivered it as the owner's answer, handed it to reviewer/arbiter, resumed the refused session (refused again) and carried the banner forward into the next task. Now it is a `safety-refusal` trigger: the raw banner is suppressed, the Claude session is cleared, the paired task closes as `claude_safety_refusal`, the room gets a short Korean notice, and refusal banners are never carried forward. Global memory gains a rule not to read thinking blocks from session transcripts (the reasoning_extraction trigger of the 2026-10-12 incident). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
5.1 KiB
Global Memory
Shared, mutable memory for every room. Keep it short: only rules every room needs.
Detailed notes live in /home/claude/EJClaw/groups/global/ops/ — read the matching file when a task touches that topic:
ops/host.md— GPU (RTX 5050,/home/claude/gpu-venv, NVML mismatch fix), tmpfs/RAM rules and watchdogs, Windows CI VM on .9 (nested KVM, resource caps), .5 Docker host anddocker5contextops/credentials.md— what is insecrets.json, git.tkrmagid.kr credential helper, Claude/Codex per-service credentials and re-login runbooksops/ejclaw-ops.md— room mode policy, channel register/deregister, safe restarts, model upgrade checklist, current versions, recent EJClaw fixesops/projects.md— javis_bot deployment target, washing_machine_app release process
User preferences
- Answer the user in Korean 존댓말 unless asked otherwise.
- GPU: ask before running GPU workloads, except for the
javis_botproject (pre-approved).
Host safety (always)
/tmpand/dev/shmare RAM (tmpfs). Never put venvs, models, datasets or other large files there; use disk paths under$HOME(TMPDIR=$HOME/.tmp).- Heavy jobs (GPU, VMs, big builds) must not run inside the bot's cgroup: launch them with
systemd-run --user --scope -p MemoryMax=... .... - Never run a plain
systemctl --user restart ejclaw.serviceinside a turn (it kills you). Room binding changes use SIGHUP; code/env changes use the detachedsystemd-runrestart. Verify before claiming success. Details:ops/ejclaw-ops.md.
Sending files
Only files under /home/claude/EJClaw/data/attachments can be attached. Copy the file there (prefer data/attachments/generated/, auto-deleted after 60 min) and reference it with a MEDIA: line. Do not widen the allowlist.
Credentials
Stored at /home/claude/.config/ejclaw/secrets.json (read with the Read tool when needed). Never paste a secret into chat or into memory files. Each entry's note explains its usage rules; see ops/credentials.md.
Rooms
- All rooms are
tribunal(owner + reviewer + arbiter) unless the user asks for single mode in that room ("리뷰어 끄자", "single로 바꿔줘"). Commands:ops/ejclaw-ops.md. - "채널 등록 해제 " means a full purge with
bun scripts/deregister-room.ts(dry-run first, DB backup first). Details:ops/ejclaw-ops.md. - A short stop-only message ("그만해", "멈춰", "중단", "stop") stops the room's running work; this is built in.
- Resuming interrupted work: rebuild context from chat messages,
conversations/, git state and visible tool results only. Never read or dumpthinkingblocks from Claude session transcripts (data/sessions/**/*.jsonl); that trips Opus safeguards (reasoning_extraction) and the whole turn is refused (incident 2026-10-12).
CLAUDE.md
Behavioral guidelines to reduce common LLM coding mistakes. Merge with project-specific instructions as needed.
Tradeoff: These guidelines bias toward caution over speed. For trivial tasks, use judgment.
1. Think Before Coding
Don't assume. Don't hide confusion. Surface tradeoffs.
Before implementing:
- State your assumptions explicitly. If uncertain, ask.
- If multiple interpretations exist, present them - don't pick silently.
- If a simpler approach exists, say so. Push back when warranted.
- If something is unclear, stop. Name what's confusing. Ask.
2. Simplicity First
Minimum code that solves the problem. Nothing speculative.
- No features beyond what was asked.
- No abstractions for single-use code.
- No "flexibility" or "configurability" that wasn't requested.
- No error handling for impossible scenarios.
- If you write 200 lines and it could be 50, rewrite it.
Ask yourself: "Would a senior engineer say this is overcomplicated?" If yes, simplify.
3. Surgical Changes
Touch only what you must. Clean up only your own mess.
When editing existing code:
- Don't "improve" adjacent code, comments, or formatting.
- Don't refactor things that aren't broken.
- Match existing style, even if you'd do it differently.
- If you notice unrelated dead code, mention it - don't delete it.
When your changes create orphans:
- Remove imports/variables/functions that YOUR changes made unused.
- Don't remove pre-existing dead code unless asked.
The test: Every changed line should trace directly to the user's request.
4. Goal-Driven Execution
Define success criteria. Loop until verified.
Transform tasks into verifiable goals:
- "Add validation" → "Write tests for invalid inputs, then make them pass"
- "Fix the bug" → "Write a test that reproduces it, then make it pass"
- "Refactor X" → "Ensure tests pass before and after"
For multi-step tasks, state a brief plan:
1. [Step] → verify: [check]
2. [Step] → verify: [check]
3. [Step] → verify: [check]
Strong success criteria let you loop independently. Weak criteria ("make it work") require constant clarification.
These guidelines are working if: fewer unnecessary changes in diffs, fewer rewrites due to overcomplication, and clarifying questions come before implementation rather than after mistakes.