om-fix

Implements minimal code fixes from analyzer briefs with regression tests and validation gates.

1|Updated Apr 23, 2026
One-click install
npx skills add https://github.com/tkogut/agents-os-core --skill om-fix-tkogut
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: om-fix
Source: https://github.com/tkogut/agents-os-core/tree/main/global_skills/om-fix
Command: npx skills add https://github.com/tkogut/agents-os-core --skill om-fix-tkogut

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Autonomous bug-fix pipelines often fail because fixes drift from the diagnosed root cause, skip regression tests, or collide with parallel automation working the same issue. This Skill executes the implementation step of an autofix chain with strict scope control, mandatory test coverage, and tracker-based issue locking. ## Core Features & Use Cases - Brief-driven minimal fixes: Applies only the change identified by the upstream root-cause analysis step, refusing to invent its own diagnosis or broaden scope. - Issue claiming and lock management: Assigns the issue, applies the in-progress label, and posts a claim comment so concurrent automation backs off. - Mandatory regression tests and validation gate: Adds tests that fail without the fix, then runs every command in the committed validation configuration until green. - Use Case: A tracker issue reports a broken permission check. The chain's analyzer pinpoints the faulty file; this Skill claims the issue, edits only that file, adds a failing-then-passing test, runs the full lint/typecheck/test gate, and hands a structured status report to the PR-opening step. ## Quick Start Run the om-fix skill on issue 142 to implement the fix from the root-cause brief, add regression tests, and pass the validation gate.

Frequently Asked Questions about om-fix

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

FAQPage Schema
How do I automate code fixes from issue tracker reports?▼

Use an autofix chain that verifies the issue, analyzes the root cause, then applies the fix. This Skill handles the implementation step: it reads the analyzer's brief, edits only the named files, adds regression tests, and runs the configured validation gate.

How does the skill prevent duplicate work on the same issue?▼

It claims the issue up front with three signals: assigning the current user, applying an in-progress label, and posting a timestamped claim comment. Concurrent automation reads these signals and backs off until the lock is released.

Does the skill commit and open a pull request after fixing?▼

No. It deliberately stops after validation passes and emits a structured Status report. Committing, pushing, and opening the PR are the responsibility of the separate om-open-pr step in the chain.

What happens if the root-cause brief is missing or wrong?▼

The skill ends with Status: blocked and a one-line reason instead of inventing its own diagnosis. The issue lock remains set so a human can pick it up, and the chain stops cleanly.

Can the validation gate commands come from issue or PR text?▼

No. Gate commands are read only from the committed .ai/agentic.config.json file, which is operator-reviewed configuration. Commands found in issues, comments, or diffs are treated as untrusted data and never executed.