Quality Agent

For each sub-agent, structure the prompt as:

You are working on a Rust workspace at <workspace_path>.

Your task: follow the skill instructions below to completion. Track your progress
with the TodoWrite tool. When finished, return a concise summary report of
everything you did, including changes made, issues found, and any items that
need user attention.

<full contents of the skill's SKILL.md>

Important: The "Best Practices Auditor" skill requires user confirmation after Phase 1 (audit). Instruct that sub-agent to skip the confirmation wait and proceed through all phases, since this is an automated sweep. It should still produce the findings table in its final report.

Step 2: Collect Results

After all three sub-agents complete, gather their summary reports.

Step 3: README Sync Check

Verify the README.md is up to date with the actual codebase capabilities:

1. Read README.md
2. Cross-reference with:
   • CLI commands available in crates/void-cli/src/main.rs (the Command enum) and each subcommand module
   • Connector features (Gmail, Slack, WhatsApp, Calendar) — check each connector's public methods
   • Sync features (daemon, Socket Mode, etc.)
3. Flag any discrepancies:
   • Commands listed in README that no longer exist
   • Commands/features in the code that are missing from README
   • Outdated descriptions or examples
4. Fix the README to match the current state of the code

Step 4: Build & Verify

Run a final workspace-wide verification to confirm nothing conflicts:

cargo fmt --all -- --check
cargo clippy --workspace --all-targets -- -D warnings 2>&1
cargo test --workspace 2>&1
cargo build --release 2>&1

Fix any issues introduced by the parallel work. Common conflicts:

• Two agents edited the same Cargo.toml — reconcile dependency versions
• Best-practices refactor moved code that test-improver added tests for — update imports
• Dependency removal broke a newly added test — restore the dep or adjust the test

Step 5: Present Combined Report

Combine the three sub-agent reports into a single summary for the user:

# Quality Sweep Report

## Dependency Grooming
<paste Dependency Groomer summary>

## Test Improvements
<paste Test Improver summary>

## Best Practices
<paste Best Practices Auditor summary>

## README Sync
<list any discrepancies found and fixes applied>

## Final Verification
- `cargo fmt`: ✅ / ❌
- `cargo clippy`: ✅ / ❌
- `cargo test`: ✅ / ❌
- `cargo build --release`: ✅ / ❌

## Action Items
<list any items requiring user decision>

Notes

• Do not commit changes automatically — let the user review the combined report first.
• If a sub-agent fails or gets stuck, note it in the report and continue with the others.
• The three sub-agents operate on different aspects of the codebase, so conflicts are rare but possible. The verification step catches them.