This release fixes issues for SREs watching stability and regressions.
✓ No known CVEs patched in this version
Topics
+14 more
Summary
AI summaryFixes ensure that zero-output streams and incomplete subagent results cause failures rather than incorrect success states.
Changes in this release
| Type | Severity | Summary | CVE |
|---|---|---|---|
| Bugfix | Medium |
Fixes subagent zero-output stream-incomplete runs to fail instead of succeed. Fixes subagent zero-output stream-incomplete runs to fail instead of succeed. Source: llm_adapter@2026-07-16 Confidence: high |
— |
| Bugfix | Medium |
Prevents Telegram cleanFinal delivery from re‑sending contradictory overflow tail on failure. Prevents Telegram cleanFinal delivery from re‑sending contradictory overflow tail on failure. Source: llm_adapter@2026-07-16 Confidence: high |
— |
| Bugfix | Medium |
Hard‑fails mint phases when subagent result is incomplete or capped. Hard‑fails mint phases when subagent result is incomplete or capped. Source: llm_adapter@2026-07-16 Confidence: high |
— |
Full changelog
What's Changed
- fix(subagent): resolve zero-output stream-incomplete run as failed, not succeeded by @griffinwork40 in https://github.com/griffinwork40/agent-afk/pull/628
- fix(telegram): failed cleanFinal delivery no longer re-sends a contradictory overflow tail (#623) by @griffinwork40 in https://github.com/griffinwork40/agent-afk/pull/626
- fix(mint): hard-fail phases on incomplete subagent result (capped / stream_incomplete) by @griffinwork40 in https://github.com/griffinwork40/agent-afk/pull/627
Full Changelog: https://github.com/griffinwork40/agent-afk/compare/v5.52.0...v5.52.1
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About Agent AFK
Coding-agent harness built for unattended, away-from-keyboard runs across four surfaces sharing one session manager: one-shot CLI, REPL, cron-friendly headless daemon, and a Telegram bot. Editable agent loop (prompts, permission gates, model routing, explicit terminal states), MCP (stdio/HTTP/SSE + OAuth), lifecycle hooks, background subagents, plan mode, cross-session memory, and an append-only trace receipt (`afk trace show`). Works with Anthropic and any OpenAI-compatible endpoint (GPT, Codex, local MLX/llama.cpp/Ollama)
Related context
Related tools
Beta — feedback welcome: [email protected]