microsoft / microsoft/agent-governance-toolkit
Implementation follow-up: EvidenceAnchor / mycelium-agt — ecosystem developments since #2244 merged
- Dominant language
- Python
- Stars
- 6.3k
- Forks
- 1.1k
- Avg merge
- 5d 11h
- Merged PRs (30d)
- 142
Description
Following up on the proposal merged in #2244 (docs/proposals/MYCELIUM-EXTERNAL-ANCHOR-PROPOSAL.md, 2026-05-18).
Two developments since the proposal landed:
1. The action_ref derivation spec is now an IETF Internet-Draft: draft-etcheverry-action-ref, currently at -03.
2. Three independent implementations have contributed cross-verified conformance sets to argentum-core: AEOESS/APS (examples/conformance/aps/), presidio-v/presidio-hardened-x402 (examples/conformance/presidio/), and TKCollective/agentoracle-receipt-spec (examples/conformance/agentoracle-v1/). Each reproduces the four-field preimage byte-identical against the spec.
The proposal is the design baseline. Happy to open the implementation PR if that's still the right path, or to follow whatever integration track makes sense given where AGT has moved since May.
---
Edit 2026-09-08: corrected the draft name above (was mistakenly written as draft-giskard-aeoess-action-ref, with "AEOESS" wrongly expanded as an organization name — see @aeoess' correction in the comments).
Contributor guide
Research direction
Read docs/proposals/MYCELIUM-EXTERNAL-ANCHOR-PROPOSAL.md first, then inspect the conformance sets under examples/conformance/aps/, examples/conformance/presidio/, and examples/conformance/agentoracle-v1/. Determine the appropriate integration track for the EvidenceAnchor and action_ref work; the issue is complete when that track and implementation scope are agreed.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- ai-infra-agents, security
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Active
- Clarity
- Needs clarification
- Newbie friendliness
- 35/100