Skip to content

chore: one agent-action-capsule and one checkpointed-local-log line - #279

Open
ethanyzhang wants to merge 4 commits into
Mesh-LLM:mainfrom
ethanyzhang:chore/aac-one-line
Open

ethanyzhang wants to merge 4 commits into
Mesh-LLM:mainfrom
ethanyzhang:chore/aac-one-line

Conversation

@ethanyzhang

Copy link
Copy Markdown
Collaborator

Stacked on #278 (this branch contains its two commits). Moves the evidence-plugin to one agent-action-capsule line (0.10.0) and one checkpointed-local-log line (0.5.0): evidencebook =0.0.15, evidence-request =0.5.0, capsule-emit 0.0.12, capsule-emit-artifact 0.0.3, cll 0.5.0 (the crates are published: evidence-request 0.5.0, evidencebook 0.0.15, capsule-emit 0.0.12, capsule-emit-artifact 0.0.3).

Code changes the new versions require, none changing behavior:

  • ChainLog implements cll 0.5.0 WitnessStateStore::pending_witnesses_for by delegating to its log, like the other methods.
  • evidencebook::answer::Options gains standard_card; the plugin passes None, which is what it had before the field existed (the standard_history derivation stays unsupported).
  • evidence-request 0.5.0 also lists evidence-interop/2, which the node does not answer; evidence_answer now gates on its own list (/0, /1) so a /2 request is still request_malformed (the existing test a_request_for_a_profile_not_implemented_is_malformed passes unchanged).

Checked in crates/evidence-plugin: cargo fmt --check, cargo clippy --locked --all-targets -- -D warnings, cargo test --locked -- --test-threads=1 (1465 passed, 32 ignored), cargo tree -d --locked shows one agent-action-capsule, one checkpointed-local-log (evidence-request twice, 0.4.1 from agent-action-capsule 0.10.0 without its cll feature). Part of action-state-group/capsule-emit#446.

Signed-off-by: Ethan Zhang <github@yzhang.io>
…evidencebook 0.0.15, evidence-request 0.5.0)

Signed-off-by: Ethan Zhang <github@yzhang.io>
@coderabbitai

coderabbitai Bot commented Nov 9, 2026 •

Copy link
Copy Markdown

Important

  • Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

Run configuration

  • Source mode: Auto
  • Repository: Unavailable
  • Branch: Unavailable
  • File: No configuration file from this repository applied
  • Commit: Unavailable
  • Layers: Organization UI
  • Review profile: CHILL
  • Plan: Advanced
  • Run ID: 1e73823a-24a5-48ef-80ed-ccbe6ee570b9
  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

Share

…omments

Signed-off-by: Ethan Zhang <github@yzhang.io>
@ethanyzhang

Copy link
Copy Markdown
Collaborator Author

(Ethan's agent) Review record, PR #279. Reviewer: one Claude stand-in (claude -p --permission-mode plan) for Codex and Kimi (Codex out of quota until 11-11), round 1; no Sherlock2 (no product rule; public text is a changelog entry stating the dependency move).

  • P1 fixed (behavior preservation): the requester side evidence_artifact::check would have accepted a valid evidence-interop/2 envelope (evidence-request 0.5.0 lists /2). ANSWERED_PROFILES (/0, /1) now gates both the responder and check; an_answer_under_another_profile_than_asked_is_refused has a /2 case.
  • P3 fixed: stale Cargo.toml comments, changelog now names the second evidence-request line and capsule-emit 0.0.12's witness-client behavior.
  • P2 not changed (exists before this PR, wire bytes unchanged): a /2 asker cannot verify this node's refusal of a /2 request because the A7-signed refusal carries policy_digest outside the signature. Follow-up if wanted.
  • Targeted check on the fixes: cargo fmt --check, cargo clippy --locked --all-targets -D warnings, cargo test --locked -- --test-threads=1 (1465 passed, 32 ignored), all exit 0 locally on this head.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant