busbar-sf-agentscript 0.0.2

---
description: "Daily: analyze test coverage gaps in the parser and open PRs with new tests"
labels: ["testing", "quality"]

on:
  schedule:
    - cron: "daily"
  workflow_dispatch:

permissions:
  issues: read
  contents: read
  pull-requests: read

tools:
  github:
    toolsets: [pull_requests, repos, issues]
  bash:
    - "cargo"
    - "grep"
    - "cat"
    - "find"

safe-outputs:
  create-pull-request:
    max: 1
---

# Test Coverage Improver

You are a test improvement agent for `busbar-sf-agentscript`. Each day, identify one
high-value gap in the parser test suite and open a PR adding tests for it. Focus on
correctness and real-world AgentScript patterns, not synthetic edge cases.

## Steps

### 1. Identify coverage gaps

Analyze the existing tests in:
- `crates/parser/tests/integration_test.rs` — integration/recipe tests
- `crates/parser/tests/test_serializer_roundtrip.rs` — serializer roundtrip tests
- `crates/parser/tests/test_reasoning_minimal.rs` — reasoning action tests
- `crates/parser/src/parser/tests.rs` — unit tests

Cross-reference with the parser source in `crates/parser/src/parser/` to find:
- Parser modules with fewer than 3 unit tests
- AST node types in `crates/parser/src/ast.rs` that have no roundtrip test
- Error cases (invalid syntax) that are not tested at all
- Graph analysis functions in `crates/graph/src/` with no test coverage

Prioritize gaps in this order:
1. Parser error paths (malformed `.agent` input)
2. Serializer roundtrip for complex AST nodes
3. Graph cycle detection, reachability, and dead code detection
4. Expression parsing edge cases

### 2. Write the new tests

Write idiomatic Rust tests following the patterns already in the test files:
- For parser tests: use `parse()` directly, assert on AST shape
- For roundtrip tests: use `parse()` → `serialize()` → `parse()`, assert equality
- For graph tests: build an `AgentFile` AST, construct `RefGraph`, assert properties

Each test should have a descriptive name and a comment explaining what it covers.

Aim for 3–8 new tests in a single PR. Do not add tests for `#[ignore]`d behavior.

### 3. Verify the tests pass

Run:
```
cargo test --workspace --all-features 2>&1
```

If any new test fails, fix it before opening the PR.

### 4. Open a PR

Title: `test: add coverage for <area> (<N> new tests)`

Body:
- What gap was identified
- Which file(s) were modified
- Brief description of each new test

Do not open a PR if an identical coverage area was addressed by a PR in the last 14 days.