This release fixes issues for SREs watching stability and regressions.
Published 20d
AI Agents & Assistants
✓ No known CVEs patched
✓ No known CVEs patched in this version
Topics
agent-runtime
ai-agent-runtime
ai-coding-agents
ai-control-plane
ai-governance
ai-safety
+13 more
audit-trail
budget-enforcement
claude-code
codex
coding-agents
control-plane
governed-runtime
llmops
mcp
model-context-protocol
observability
policy-as-code
rollback
Summary
AI summaryRestored consistent error normalization with parseable failure messages and improved Codex launch diagnostics.
Full changelog
Fixed
- Error normalization restored —
failure.messagenow consistently carries the normalized, parseable failure signal that downstream consumers depend on. Raw stderr is preserved in thesummaryfield for diagnostics. Regression introduced in 0.4.1. - Improved Codex launch failure diagnostics — When Codex exits immediately with no output, the failure summary now surfaces the actual exit code and stderr content instead of a generic message, making failures actionable.
- Token flag regression guard —
--max-tokens(which does not exist in the Claude CLI) is now explicitly tested as absent. Token enforcement is via the streaming budget circuit breaker.
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 Keesan12/Martin-Loop
All releases →Related context
Related tools
Earlier breaking changes
- vmcp-v0.1.3 martin_status uses oneOf for selector exclusivity, latest as const.
- vmcp-v0.1.3 maxIterations and maxTokens modeled as integers in tool schemas.
- vmcp-v0.1.3 Tool schemas enforce additionalProperties: false on public contracts.
- vmcp-v0.1.3 Packaged artifacts now require and ship server.json alongside package.json.
Beta — feedback welcome: [email protected]