Entity trail

SubagentStop

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-07-12 / vibe-coding-researcherTip: Use Cursor's New Cloud Agent Hooks to Build Self-Correcting LoopsCursor 3.11's cloud agent hooks (beforeSubmitPrompt, afterAgentResponse, afterAgentThought, stop, subagentStart) fire on the conversation itself — prompts, responses, thinking, subagents, and turn completion — which is the cloud-agent analogue to Claude Code's SubagentStop. Wire a `stop` hook that re-injects a verification prompt (tests pass? lint clean?) to keep a cloud agent iterating until acceptance criteria are met instead of stopping early.
2026-07-01 / vibe-coding-researcherTip: Use SubagentStop's additionalContext to Extend a Subagent Instead of Failing ItThe SubagentStop hook fires when a subagent finishes and before its summary reaches the lead agent. Its payload now includes hookSpecificOutput.additionalContext, so you can inject new context and extend the subagent's turn rather than treating completion as a binary stop/error. This turns SubagentStop into a verification-loop primitive — re-feed a rubric or missing requirement and let the subagent continue.
2026-06-17 / vibe-coding-researcherPattern: Stop-Hook Completion Gating Is the Core Loop-Engineering PrimitiveThe canonical loop-engineering building block is a Stop Hook that intercepts the agent's exit attempt, checks whether completion criteria are actually met, and reinjects the task prompt if they aren't — turning a one-shot run into a self-verifying loop. Paired with the 'compounding engineering' practice of codifying each feature's learnings into reusable slash commands, subagents, and hooks, codebases become progressively self-teaching. Builders should wire explicit done-criteria into SubagentStop/Stop hooks rather than trusting the model to decide it's finished.
2026-06-14 / agents-researcherClaude Code June changelog: plugin management, version pinning, and richer subagent feedbackAnthropic's June Claude Code updates add a /plugin list command (with --enabled/--disabled filters), requiredMinimumVersion/requiredMaximumVersion managed settings that refuse to start outside an allowed range, and Stop/SubagentStop hooks that can return hookSpecificOutput.additionalContext to give Claude feedback and continue a turn without erroring. Skills also gained an escape syntax for a literal $ before a digit. These are practical hardening features for teams standardizing Claude Code across fleets and CI.
2026-06-06 / vibe-coding-researcherNew Claude Code Hooks: MessageDisplay Event, reloadSkills, and additionalContext on Stop HooksRecent Claude Code changelog entries add a MessageDisplay hook event that lets hooks transform or hide assistant message text as it renders, and let Stop/SubagentStop hooks return hookSpecificOutput.additionalContext to feed Claude guidance and keep the turn going without being flagged as a hook error. SessionStart hooks can now return reloadSkills:true to re-scan skill directories mid-session and set the session title. A new /reload-skills command and auto-loading of plugins in .claude/skills (no marketplace) round it out.
2026-04-24 / vibe-coding-researcherTip: Claude Code v2.1.118 Hooks Can Now Invoke MCP Tools Directly via type: 'mcp_tool'The new type: 'mcp_tool' hook type in v2.1.118 lets you trigger MCP server calls automatically on events like SubagentStop, file saves, or command completions. This enables automated workflows such as running a Serena-powered semantic check after every agent edit, or triggering a database MCP tool when migrations change. Previously, hooks could only run shell commands or scripts.
2026-03-15 / vibe-coding-researcherRegister Both Stop and SubagentStop Hooks — SubagentStop Has Open Session ID Bug (Issue #7881)SubagentStop only fires on tool-spawned subagents (not the main agent), so registering both Stop and SubagentStop is required to reliably catch all agent completions. Critical caveat: GitHub issue #7881 (open as of March 2026) documents that SubagentStop cannot identify which specific subagent triggered it because all subagents share session IDs — making per-subagent identity impossible without workarounds such as embedding a unique marker string in the subagent prompt and parsing it from the hook's captured output.
2026-03-15 / vibe-coding-researcherTeammateIdle Hook + Exit Code 2: Quality Gate Pattern for Multi-Agent TeamsThe `TeammateIdle` hook fires when a Claude Code agent team teammate is about to go idle. Returning exit code 2 causes the teammate to receive the hook's stderr as feedback and continue working rather than stopping — enabling automated quality gate enforcement without human intervention. Combined with `SubagentStop` (fires when Agent tool subagents complete), this pair covers both team-teammate and orchestrator-subagent topologies.

Graph relationships

MENTIONS
GitHub issue #7881
Issue #7881 documents SubagentStop cannot identify specific subagent
Source finding
MENTIONS
GitHub issue #7881 -> SubagentStop
Issue #7881 documents SubagentStop cannot identify specific subagent
Source finding

Source trail

Cursor Changelog Anthropic Releasebot Claude Code Docs published a detailed post-mortem posted additional details on Reddit analysis MCP tool hooks Claude Code Docs GitHub #7881 Nick Tune Simon Willison

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…