Parent roadmap: #82
Background
MCP users need short, useful explanations they can act on immediately, not just scores or labels.
Goal
Add compact explanation cards to agent outputs that state why an action helps now and when to rerun.
Current Behavior
Some reasoning exists, but it can be verbose, buried, or not tied to specific next actions.
Desired Behavior
Each action includes a concise explanation card with scoreability blocker, risk, maintainer friction, expected impact, and rerun condition.
Implementation Requirements
- Add explanation card fields to agent action output.
- Keep copy deterministic and structured.
- Separate branch blockers from account/queue blockers.
- Include rerunWhen.
- Preserve JSON-first MCP output.
Public/Private Output Boundaries
Private MCP/API may include scoreability, risk, and confidence. Public GitHub output must use sanitized summaries only and must not expose private scoring, reward/risk, raw trust, wallet/hotkey, or reviewability internals.
Acceptance Criteria
- Cards are present for all top actions.
- Cards explain cleanup/wait/pursue/avoid differently.
- Cards are deterministic for same input.
- Public-safe variants are sanitized.
Testing Requirements
npm run test:ci must pass.
- Global coverage must remain at or above 97% for lines, statements, functions, and branches.
- Aim for 98%+ branch coverage locally to avoid CI variance.
- Add tests for every new branch, fallback path, sanitizer rule, and regression.
- Add invariant/property-style tests when behavior depends on sorting, gating, scoring, queue pressure, source-upload safety, public/private boundaries, or upstream drift.
- Public GitHub output must be tested against forbidden language: wallet, hotkey, raw trust score, payout, reward estimate, farming, private reviewability, and public score estimate.
- MCP/local tooling must prove source contents are not uploaded.
Additional Test Scenarios
- Snapshot tests for card text shape.
- Branch blocker fixture.
- Account blocker fixture.
- Public sanitizer regression.
Parent roadmap: #82
Background
MCP users need short, useful explanations they can act on immediately, not just scores or labels.
Goal
Add compact explanation cards to agent outputs that state why an action helps now and when to rerun.
Current Behavior
Some reasoning exists, but it can be verbose, buried, or not tied to specific next actions.
Desired Behavior
Each action includes a concise explanation card with scoreability blocker, risk, maintainer friction, expected impact, and rerun condition.
Implementation Requirements
Public/Private Output Boundaries
Private MCP/API may include scoreability, risk, and confidence. Public GitHub output must use sanitized summaries only and must not expose private scoring, reward/risk, raw trust, wallet/hotkey, or reviewability internals.
Acceptance Criteria
Testing Requirements
npm run test:cimust pass.Additional Test Scenarios