Skip to content

Latest commit

 

History

History
83 lines (60 loc) · 4.05 KB

File metadata and controls

83 lines (60 loc) · 4.05 KB

Engineering Loops: Standardized Workflows

Domain: AI / Workflows
Status: Canonical

Summary

Engineering Loops are the standardized execution plans used by agents to resolve Drift. By following these deterministic sequences, dev.kit ensures that complex tasks—from feature implementation to documentation synchronization—remain grounded in repository truth.


🏗 The Standard Loop (Drift Resolution)

Every high-fidelity task follows the Analyze -> Normalize -> Process -> Validate -> Capture cycle.

1. Feature Engineering Loop

Standard loop for implementing new capabilities with TDD and documentation.

  • Goal: Expand repository "Skills" while maintaining 12-factor compliance.
  • Steps:
    1. Analyze: Audit existing code and docs to identify the implementation gap.
    2. Normalize: dev.kit task start to create a bounded workflow.md.
    3. Process: dev.kit skills run to implement logic and test cases.
    4. Validate: dev.kit doctor to verify environment health and TDD success.
    5. Capture: dev.kit sync run to logically group and commit the resolution.

2. Resilient Bugfix Loop

Deterministic lifecycle for identifying, reproducing, and resolving repository defects.

  • Goal: Restore repository integrity with verified test evidence.
  • Steps:
    1. Analyze: dev.kit doctor to detect environment or software drift.
    2. Normalize: Define reproduction steps in a new workflow.md.
    3. Process: Apply the fix and implement a regression test.
    4. Validate: Execute the test suite within the Worker Ecosystem.
    5. Capture: dev.kit sync run to finalize the fix and update the Skill Mesh.

3. Knowledge & Discovery Sync

Workflow for synchronizing repository documentation and agent context.

  • Goal: Eliminate documentation drift and hydrate the Skill Mesh.
  • Steps:
    1. Analyze: Scan docs/ and script headers for outdated metadata.
    2. Normalize: Map documentation updates to current repository reality.
    3. Process: dev.kit visualizer to regenerate high-fidelity architecture diagrams.
    4. Validate: Verify that all internal and external links are high-fidelity.
    5. Capture: dev.kit ai sync to ground the agent in the updated knowledge.

🏗 Standard Loop Mapping

The standard engineering loops are operationalized through specialized UDX targets:

Loop Domain Grounding Target Pattern Role
Logic Implementation udx/dev.kit Intent normalization and task management.
Environment Parity udx/worker Deterministic runtime for loop execution.
Automation Flow udx/reusable-workflows Pattern baseline for sequence steps.

🏗 Workflow Grounding

Engineering loops are operationalized through deterministic UDX engines:

Loop Type Grounding Resource Role
Engineering udx/dev.kit Dynamic normalization and task management.
Automation udx/reusable-workflows Pattern baseline for implementation steps.

🧠 Continuity Mandates

  • Resume First: Before starting a new loop, agents must check for active tasks (dev.kit task active).
  • Hygiene: Aborted or stagnant loops must be pruned (dev.kit task cleanup) to prevent context noise.
  • Feedback: Every iteration must emit high-signal progress to the feedback.md artifact.

📚 Authoritative References

Standardized loops ensure predictable delivery and high-fidelity results:


UDX DevSecOps Team