chore(claude): allowlist persona-sati MCP tools to prevent non-interactive REJECT#11
Merged
Merged
Conversation
…ctive REJECT Adds a PreToolUse hook with matcher ^mcp__persona-sati__ that returns permissionDecision=allow. Prevents Claude Code from re-prompting for persona-sati tool permissions during long sessions, which on 2026-05-28 hung the main LLM thread for 3h 21m when a re-prompt fired while the user was AFK and the default-deny path triggered. Server-side fixes for related issues already shipped on persona-sati main (PRs #56, #61, #62). Full incident analysis: persona-sati hard-won-learnings Learning #39. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Adds a
PreToolUsehook in.claude/settings.jsonwith matcher^mcp__persona-sati__that returnspermissionDecision: allow. Bypasses Claude Code's permission gate for any persona-sati tool call so a non-interactive permission re-prompt can never fire.Why
On 2026-05-28, persona-sati's MCP session produced a 3h 21m hang in the entrabot main LLM thread. Root cause: Claude Code's permission gate fired a re-prompt on
mcp__persona-sati__observewhile the user was away from the terminal, and the default-deny-on-no-human path produced aREJECTEDstatus. The agent loop then entered a stuck reconciliation state.Persona-sati server-side fixes for related issues already shipped on persona-sati main (PRs #56 bound sync I/O, #61 structured logging, #62 Container Insights opt-in). This is the client-side companion change — together they belt-and-suspenders the failure mode.
Full incident write-up: persona-sati
docs/runbooks/hard-won-learnings.mdLearning #39.Behavior change
Before: persona-sati tool calls may trigger a Claude Code permission prompt that, in non-interactive sessions (user AFK), defaults to deny → REJECTED status → main thread can hang in reconciliation.
After: persona-sati tools are silently allowed by the hook. No permission prompt, no REJECTED outcome.
Existing hooks (
Write/Edit/NotebookEdit, entrabot send-side tools, SessionStart, PostToolUse) are untouched.Test plan
python3 -c "import json; json.load(open('.claude/settings.json'))") — confirmed locally before commit🤖 Generated with Claude Code