feat(skills): expose per-expectation effect outcomes

[yuecideng]

Description

Stack

• Layer: 22/28
• Base: feat/declarative-hand-over-task
• Depends on: feat(tasks): add declarative physical handover #505
• Followed by: feat(atomic-actions): reconcile terminal effect failures #478

Expose immutable, row-local outcomes for every physical effect expectation. The composite monitor now maintains hysteresis independently per expectation and reports satisfied, contradicted, and fully inverse-satisfied masks from one observation, while preserving aggregate success/failure masks as exact reductions.

This gives later failure reconciliation enough evidence to distinguish a single contradictory clause from complete inverse evidence. It also aligns the cube and drawer test robot fixtures with the existing get_qpos(target=...) simulation contract so every layer in this stack is independently testable.

Refs #471
Refs #474

Type of change

• New feature (non-breaking change which adds functionality)

Screenshots

Not applicable.

Validation

• pytest -q tests/sim/skills/test_effects.py tests/gym/envs/tasks/test_open_drawer.py tests/gym/envs/tasks/test_multi_segments_cube_pick_place.py — 43 passed, 1 deselected
• Black 26.3.1 check on changed Python files

Checklist

• Changed Python files pass Black 26.3.1.
• Public documentation is completed by follow-up layer test(tasks): gate cube physical recovery #483.
• Tests cover the new behavior.
• No dependency update is required.

[greptile-apps]

Greptile Summary

This PR adds immutable per-expectation effect outcomes and changes composite monitoring to maintain hysteresis independently for each expectation while deriving aggregate success and failure from the current observation.

• Exports EffectExpectationDecision with satisfied, contradicted, and inverse-satisfied masks.
• Tracks success, contradiction, and complete inverse evidence per expectation and environment.
• Adds transfer-monitor coverage and updates simulation fixtures to accept get_qpos(target=...).

Confidence Score: 4/5

The implementation appears safe to merge from a runtime perspective, with the non-blocking omission that its newly exported public API is not documented in this layer.

The per-expectation reductions and hysteresis paths have focused tests and no concrete behavioral failure remains, but users lack public documentation for the newly exported decision contract.

Files Needing Attention: embodichain/lab/sim/skills/effects.py and embodichain/lab/sim/skills/init.py

Important Files Changed

Filename	Overview
embodichain/lab/sim/skills/effects.py	Adds immutable per-expectation outcomes and independent hysteresis, but the newly public API is not documented in this PR.
embodichain/lab/sim/skills/init.py	Re-exports EffectExpectationDecision as part of the public skills API.
tests/sim/skills/test_effects.py	Adds focused coverage for mask ownership, inverse evidence, per-expectation hysteresis, and prevention of cross-tick success stitching.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A[Evidence observation] --> B[Classify clauses by expectation]
    B --> C[Update row-local hysteresis counters]
    C --> D[Build satisfied mask]
    C --> E[Build contradicted mask]
    C --> F[Build inverse-satisfied mask]
    D --> G[AND all satisfied masks]
    E --> H[OR all contradicted masks]
    G --> I[EffectMonitorDecision]
    H --> I
    D --> J[Per-expectation decisions]
    E --> J
    F --> J
    J --> I

Loading

Prompt To Fix All With AI

### Issue 1
embodichain/lab/sim/skills/effects.py:1513-1514
**Document per-expectation outcomes**

`EffectExpectationDecision` and `EffectMonitorDecision.expectation_decisions` are newly exported public APIs, but this layer does not document their mask semantics or aggregate relationship, making the new contract undiscoverable through the public documentation.

---

For each issue above, determine whether it is valid and should be fixed. If so, fix it directly.

_{Reviews (1): Last reviewed commit: "feat(skills): expose per-expectation eff..." | Re-trigger Greptile}

[greptile-apps]

Document per-expectation outcomes

EffectExpectationDecision and EffectMonitorDecision.expectation_decisions are newly exported public APIs, but this layer does not document their mask semantics or aggregate relationship, making the new contract undiscoverable through the public documentation.

Context Used: AGENTS.md (source)

Prompt To Fix With AI

This is a comment left during a code review.
Path: embodichain/lab/sim/skills/effects.py
Line: 1513-1514

Comment:
**Document per-expectation outcomes**

`EffectExpectationDecision` and `EffectMonitorDecision.expectation_decisions` are newly exported public APIs, but this layer does not document their mask semantics or aggregate relationship, making the new contract undiscoverable through the public documentation.

**Context Used:** AGENTS.md ([source](https://github.com/dexforce/embodichain/blob/main/AGENTS.md))

---

For each issue above, determine whether it is valid and should be fixed. If so, fix it directly.

Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!

[yuecideng]

Folded into #480 during stacked-PR consolidation. Its commits remain included in #480; the remote branch is retained for traceability.