diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md index bd74df1b..2f411983 100644 --- a/CONTRIBUTING.md +++ b/CONTRIBUTING.md @@ -8,7 +8,7 @@ Search existing issues and pull requests to avoid duplicating efforts. ## Current Priority Scope -Reviewed at each milestone. Current window: the open milestones on the [milestones page](https://github.com/GenAI-Security-Project/agent-control-standard/milestones). Next review October 9, 2026. +Reviewed at each milestone. Current window: the open milestones on the [milestones page](https://github.com/GenAI-Security-Project/agent-control-standard/milestones). Next review December 9, 2026. The project has one committed outcome: @@ -24,16 +24,16 @@ or a comment, this wins. **In focus.** Work that feeds the runnable Guardian, the interoperability benchmark, or the conformance evidence that makes either credible: -- The mandatory floor decision in PR #21, which gates everything behind it - The Claude Code, Cursor, and NVIDIA NAT adapters in PR #22 - The AGT reference implementation in `reference-implementations/agt/` - Ports of that reference implementation to other runtimes: Python, Go, Rust, Codex - Production-hardening it, including span batching and OpenTelemetry collection -- Resolving the fail-open default, which issues #32 and #37 attack from opposite ends -- The ACS-Core conformance claim template -- The requirement ledger and behavioral tests in milestone #33 -- Closing the Cursor file-read gap +- Resolving the fail-open default in issue #32 +- Conformance claims against the ACS-Core claim template, issues #106 and #136 +- The requirement ledger and behavioral tests in issue #33 +- Closing the Cursor file-read gap in issue #212 - Conformance and dogfooding reports that document where ACS fails in practice +- The public site, when a change helps an evaluator find, install, or run the Guardian **Deferred to v0.2.0.** Real work, tracked, landing in a later release: async and composition, streaming, batching semantics, recursive ask, quorum, multi-tenant isolation, the Cedar @@ -180,7 +180,7 @@ shipping. ## Reporting Security Issues -**Do not file public issues for security vulnerabilities.** Use GitHub's [private vulnerability reporting](https://github.com/GenAI-Security-Project/agent-control-standard/security/advisories/new) to disclose privately. We'll acknowledge within 72 hours and coordinate a fix and disclosure timeline with you. +**Do not file public issues for security vulnerabilities.** Use GitHub's [private vulnerability reporting](https://github.com/GenAI-Security-Project/agent-control-standard/security/advisories/new) to disclose privately. [SECURITY.md](./SECURITY.md) sets the response times, and we coordinate a fix and disclosure timeline with you. ## Developer's Certificate of Origin 1.1