Spec Loop Closer

Spec Loop Closer

Prerequisites & Dependencies

• Python 3.10+ (for assertion mapping logic)
• Access to project source files and test directories
• No external runtime dependencies

Execution Steps

1. Materialize requirements: Parse the user's original request/task description and extract each distinct requirement/acceptance criterion.

2. Map to checkable assertions: For each requirement, define a machine-verifiable check:

   • Grep-able token: A unique string that can be found with grep in the implemented code
   • Executable test: A pytest/unit test that passes/fails based on the requirement
   • Numeric threshold: A specific value that can be computed and compared (e.g., "response time < 200ms")
   • Enum/constraint check: A field that must match one of a set of allowed values

3. Create assertion mapping table: Produce a structured table listing:

   • Requirement ID (auto-generated or user-assigned)
   • Original requirement text
   • Assertion type (grep/test/threshold/enum)
   • Concrete check expression or test code
   • Status (mapped / pending / blocked)

4. Block until all mapped: Implementation must not report "done" until every requirement has a corresponding mapped assertion with a machine-runnable check.

5. Post-implementation verification: Run each assertion manually or via automation to confirm all pass before marking the task complete.

Assertion Mapping Table Example

Windows PowerShell Notes

• Use Select-String instead of grep for Windows pattern matching
• Numeric threshold checks should use [regex]::Match() for PowerShell parsing
• Assertion mapping table can be saved as spec-assertions.json for machine consumption