Corrections
Corrections Log
Section titled “Corrections Log”Read this before starting work. These are real mistakes that cost time/money. (Includes legacy items from EliteBook era).
File Paths
Section titled “File Paths”- Never use
~in tool calls — it doesn’t resolve. Always use full paths:/home/chrisr6/.openclaw/workspace/...
Agent Comms
Section titled “Agent Comms”- Always
timeoutSeconds=0onsessions_send— nested lane hasconcurrency=1. Without fire-and-forget, calls queue and timeout at 30s even though the message was delivered. sessions_listalwaysmessageLimit: 0— message payloads include thinking blocks + tool calls = 20K+ tokens per call. Burned 111K context in 2hrs from this.
Provider Auth
Section titled “Provider Auth”- Auth:yes ≠ quota available — always test-complete before assigning. The
anthropic-directprovider leaked ~$15 API credits in 2 days before we caught it. - Never add unknown keys to openclaw.json — causes gateway crash. Validate with
openclaw gateway statusbefore restart.
Config Safety
Section titled “Config Safety”- Never restart gateway mid-conversation — drops all active sessions including Chris’s chat.
Sub-agents
Section titled “Sub-agents”- Never go silent after sub-agent completes — always QA results and proactively message Chris. NO_REPLY after a completion = broken workflow.
- Don’t busy-poll sub-agents — they auto-announce on completion. Check only on-demand.
- Debug loops → spawn out — if past the first fix attempt, spawn a dev sub-agent. Don’t burn orchestrator context on read→test→fix→test cycles.
Skill creation done inline instead of delegated
Section titled “Skill creation done inline instead of delegated”What happened: Created systematic-debugging skill file inline on Opus, burning ~13k tokens. Why it was wrong: Single-file creation with clear spec = textbook sub-agent task. Rule: Any file creation/editing task with a clear spec → spawn a sub-agent. Tila writes the brief, sub-agent writes the files.
Claude Code run via exec instead of sub-agent
Section titled “Claude Code run via exec instead of sub-agent”What happened: Ran claude -p --model opus via raw exec to build the horse profile scraper. Burned context, no auto-announce.
Rule: ALL Claude Code / dev work → sessions_spawn(agentId="dev", task="...", model="anthropic/claude-opus-4-6"). Never raw exec for coding tasks.
Coding sub-agents spawned without CLI tools
Section titled “Coding sub-agents spawned without CLI tools”Rule: ALL coding sub-agents MUST invoke a CLI tool — priority order: Codex (free) → Gemini CLI (free) → Claude Code (paid, last resort).