ordo-error-router

Routes build gate failures into bounded autonomous retries or hard halts.

Updated Aug 10, 2026
One-click install
npx skills add https://github.com/Rekonquest/ordo-pro --skill ordo-error-router-rekonquest
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ordo-error-router
Source: https://github.com/Rekonquest/ordo-pro/tree/main/default-user-files/skills/ordo-error-router
Command: npx skills add https://github.com/Rekonquest/ordo-pro --skill ordo-error-router-rekonquest

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When a build pipeline step gate fails, models either dump raw error traces or thrash through unbounded fixes that compile but violate the architecture. This Skill replaces both failure modes with a disciplined router that decides, for every gate failure, whether the fix is bounded and nameable (autonomous retry) or unbounded and doctrine-violating (hard halt to the user). ## Core Features & Use Cases - Bounded vs. unbounded routing: Classifies failures by whether a named, contract-preserving fix exists, not by which subsystem complained. - Capped autonomous Debugger loop: Runs at most 3 retries, feeds the full batched cargo check output to the Debugger, re-gates every retry against the same gate, and logs each attempt to a retry ledger. - Hard halt with a trail: Suspends the build, names the violated doctrine, and surfaces the retry history so the user decides next steps. - Use Case: A Rust build step fails its gate with a borrow-checker error whose only compiling fix would leak Arcs and break bus discipline. The router recognizes this as unbounded thrash, halts the build, and reports the violated doctrine instead of letting the loop ship working-but-wrong code. ## Quick Start When a step gate returns Fail, invoke the ordo-error-router skill to classify the failure and either run the capped Debugger retry loop or halt with the full error trail.

Frequently Asked Questions about ordo-error-router

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I handle build gate failures in an automated pipeline?▼

Route each gate failure by whether a bounded, nameable, contract-preserving fix exists. Bounded failures enter a capped autonomous retry loop of at most 3 attempts; unbounded or doctrine-violating failures hard-halt and surface the error trace and violated doctrine to the user.

How do I stop an AI coding loop from thrashing on borrow-checker errors?▼

Treat borrow and lifetime errors whose only compiling fix distorts ownership as unbounded failures, not bounded ones. Cap autonomous retries at 3, re-gate every retry against the same architectural gate, and halt with the retry trail when the cap is exhausted.

Should compiler errors and architectural errors be routed differently?▼

No. The correct routing axis is bounded versus unbounded, not compiler versus architectural. Some compiler errors are unbounded thrash and some architectural errors are one-line fixes, so classify by whether a named, contract-preserving fix exists.

What happens when the autonomous_correction flag is disabled?▼

When autonomous_correction is false, every gate failure hard-halts and surfaces to the user with no autonomous retries. The routing logic stays identical; the flag only controls whether the autonomous Debugger branch is live.

Why does a fix that compiles still count as a failed retry?▼

A retry that compiles but trips the step's architectural or anti-stub gate is a failed retry, not a success. This prevents clone-everything or stub-it-out fixes from sneaking a green build past the gate while violating the architecture.