Agent-fox

v4.2.1 Breaking

This release includes breaking changes for platform teams planning a safe upgrade.

Published 12d AI Agents & Assistants

View tool

✓ No known CVEs patched

Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

agentic-ai agentic-workflow

ReleasePort's take

Light signal

editorial:auto 11d

Release v4.2.1 adds the `--no-parallel` CLI flag for serial execution and refactors several APIs and documentation files.

Why it matters: The new `--no-parallel` flag lets developers enforce sequential task runs, while API and docs refactorings improve consistency and reduce maintenance overhead across impacted surfaces.

Summary

AI summary

Updates Refactors, Infrastructure, and Bug Fixes across a mixed release.

Changes in this release

Type	Severity	Summary	CVE
Feature
Feature	Medium	Adds `--no-parallel` CLI flag to force serial execution Adds `--no-parallel` CLI flag to force serial execution Source: llm_adapter@2026-07-15 Confidence: high	—
Feature	Medium	Implements task-group-scoped context rendering in `assemble_context()` Implements task-group-scoped context rendering in `assemble_context()` Source: llm_adapter@2026-07-15 Confidence: high	—
Feature	Medium	Injects preflight gate results into coder prompt as `## Preflight State` section Injects preflight gate results into coder prompt as `## Preflight State` section Source: llm_adapter@2026-07-15 Confidence: high	—
Feature	Medium	Shows subtask completion counts `(N/M)` in scoped summaries for non‑target task groups Shows subtask completion counts `(N/M)` in scoped summaries for non‑target task groups Source: llm_adapter@2026-07-15 Confidence: high	—
Bugfix	Medium	Fixes 3 post‑refactoring test failures from afissues migration Fixes 3 post‑refactoring test failures from afissues migration Source: llm_adapter@2026-07-15 Confidence: high	—
Bugfix	Low	Fixes stale `agentfox.platform` references in test docstrings Fixes stale `agentfox.platform` references in test docstrings Source: granite4.1:30b@2026-07-15-audit Confidence: high	—
Refactor
Refactor	Medium	Introduces `PreflightResult` dataclass; `run_preflight()` now returns structured gate states instead of a bare enum Introduces `PreflightResult` dataclass; `run_preflight()` now returns structured gate states instead of a bare enum Source: llm_adapter@2026-07-15 Confidence: high	—
Refactor	Medium	Adds scoped render functions `render_requirements_scoped()`, `render_test_spec_scoped()`, `render_tasks_scoped()`, and `render_individual_scoped()` to afspec Adds scoped render functions `render_requirements_scoped()`, `render_test_spec_scoped()`, `render_tasks_scoped()`, and `render_individual_scoped()` to afspec Source: llm_adapter@2026-07-15 Confidence: high	—
Refactor	Low	Removes duplicated rules from `maintainer.md`, `coder_fix.md`, and `maintainer_fix-triage.md` covered by `agent.md` Removes duplicated rules from `maintainer.md`, `coder_fix.md`, and `maintainer_fix-triage.md` covered by `agent.md` Source: llm_adapter@2026-07-15 Confidence: high	—
Refactor	Low	Rewrites "Understand Before You Code" section in `agent.md` to concise acknowledgment and guidance Rewrites "Understand Before You Code" section in `agent.md` to concise acknowledgment and guidance Source: llm_adapter@2026-07-15 Confidence: high	—
Refactor	Low	Cleans up `coder.md` by removing duplicated stale‑reference warning present in `agent.md` Cleans up `coder.md` by removing duplicated stale‑reference warning present in `agent.md` Source: llm_adapter@2026-07-15 Confidence: high	—

Full changelog

What's Changed

Features

--no-parallel CLI flag — af code --no-parallel forces serial execution (parallel=1) for a single run without editing config.toml (#716)
Task-group-scoped context rendering — assemble_context() now filters requirements, test specs, and tasks to only content relevant to the active task group, reducing prompt size by an estimated 50-65% for complex specs (#717, #720)
Preflight gate results in coder prompt — The orchestrator's preflight check results are now passed to the coder's task prompt as a ## Preflight State section, allowing the coder to skip redundant Quick Triage checks and save 2-5 tool calls per session (#718)
Subtask completion counts in scoped summaries — Non-target task groups in scoped rendering now show (N/M subtasks done) for quick orientation (#720)

Refactors

Profile deduplication — Removed duplicated rules (conventional commits, Co-Authored-By, stale references, never-push) from maintainer.md, coder_fix.md, and maintainer_fix-triage.md that were already covered by agent.md (#719)
agent.md rewrite — Replaced the "Understand Before You Code" section with a concise "Pre-Injected Context" acknowledgment and lightweight "Orient Yourself" guidance, eliminating redundant file-reading instructions (#717)
coder.md cleanup — Removed duplicated stale-reference warning and scope priority already present in agent.md (#717)

Infrastructure

PreflightResult dataclass — run_preflight() now returns a structured result with gate states (checkboxes_done, has_findings, tests_passed) instead of a bare enum (#718)
Scoped render functions — render_requirements_scoped(), render_test_spec_scoped(), render_tasks_scoped(), and render_individual_scoped() added to afspec (#717)

Bug Fixes

Resolved 3 post-refactoring test failures from afissues migration
Fixed stale agentfox.platform references in test docstrings

Full Changelog: https://github.com/agent-fox-dev/agent-fox/compare/v4.2.0...v4.2.1

View diff on GitHub

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

Share on X Share on Bluesky

Track Agent-fox

Get notified when new releases ship.

About Agent-fox

All releases →

Related context

Related tools

Earlier breaking changes

v3.6.4 Remove deprecated coding field from default config template.
v3.6.2 Removes fallback_model option from session pipeline