Problem
Helix-AGI separates semantic familiarity from exact prior-action state: its interaction ledger records stable artifact identifiers (message, email, post, comment, thread, or notification IDs) when a response action occurs, then annotates later reads when the same artifact has already been handled. Semantic similarity alone cannot answer whether an agent has already acted on this exact artifact.
Perseus has durable memory, external references, context decisions, and Ledger receipts, but this comparison did not identify a single reusable contract that lets integrations ask whether a specific external artifact was already acted on without treating that fact as ordinary semantic memory.
Proposed change
Define an optional exact-artifact interaction projection that composes Vault external references with Ledger receipts/action outcomes. It should support:
- stable artifact references and source-system/type metadata;
- prior-action lookup keyed by exact artifact identity, not substring text;
- distinct states for handled, attempted/failed, cancelled, unknown, and superseded artifact versions;
- compact read-time annotations or pre-action gates;
- scope, actor, destination, and receipt binding.
Do not copy Helix implementation code; implement the contract cleanly against canonical Perseus interfaces.
Acceptance criteria
- Exact artifact A that was acted on is distinguishable from text-identical artifact B with a different ID.
- A prior failed/cancelled action is not reported as completed.
- A changed artifact version can require a new action while preserving the prior receipt.
- Lookup is workspace/agent/destination scoped and cannot cross scopes.
- The result is machine-readable and hash-bound where it enters an action decision; raw message bodies and secrets are excluded from durable evidence.
- At least one integration fixture covers read → prior-action lookup → duplicate-action prevention → new-version allowance.
Grounding
Problem
Helix-AGI separates semantic familiarity from exact prior-action state: its interaction ledger records stable artifact identifiers (message, email, post, comment, thread, or notification IDs) when a response action occurs, then annotates later reads when the same artifact has already been handled. Semantic similarity alone cannot answer whether an agent has already acted on this exact artifact.
Perseus has durable memory, external references, context decisions, and Ledger receipts, but this comparison did not identify a single reusable contract that lets integrations ask whether a specific external artifact was already acted on without treating that fact as ordinary semantic memory.
Proposed change
Define an optional exact-artifact interaction projection that composes Vault external references with Ledger receipts/action outcomes. It should support:
Do not copy Helix implementation code; implement the contract cleanly against canonical Perseus interfaces.
Acceptance criteria
Grounding
core/interaction_ledger.py: https://github.com/munch2u-a11y/Helix-AGI/blob/main/core/interaction_ledger.py