DEV Community

Finley Sun profile picture

Finley Sun

Software engineer by day, open source contributor by night.

Location Beijing, China Joined Joined on 
A Quarantine Ledger for Agent Patches

A Quarantine Ledger for Agent Patches

Comments
6 min read
Kill Mutants in the Agent's Diff

Kill Mutants in the Agent's Diff

Comments
5 min read
Stamp the Verdict Card on the First Run

Stamp the Verdict Card on the First Run

Comments
7 min read
Freeze Flakes Before Properties Vote

Freeze Flakes Before Properties Vote

Comments
7 min read
Hold Merge Until the Trace Score Clears

Hold Merge Until the Trace Score Clears

Comments
7 min read
Destroy the Host After the Agent Verdict

Destroy the Host After the Agent Verdict

Comments
7 min read
Expire the Flake Lease Before Merge

Expire the Flake Lease Before Merge

Comments
9 min read
Seal the Fixture Before the Agent Patch

Seal the Fixture Before the Agent Patch

Comments
6 min read
The Patch Cannot Widen the Tool Surface

The Patch Cannot Widen the Tool Surface

Comments
6 min read
Record, Freeze, Then Kill the Mutant

Record, Freeze, Then Kill the Mutant

Comments
5 min read
Property Checks After the Flake Freeze

Property Checks After the Flake Freeze

Comments
6 min read
The Agent Cannot Own the Oracle

The Agent Cannot Own the Oracle

Comments
6 min read
Vacuous Greens Block the Agent Merge

Vacuous Greens Block the Agent Merge

Comments
6 min read
Shadow the Old Binary Before the Merge

Shadow the Old Binary Before the Merge

Comments
5 min read
Pin Fixture Hashes Before the Agent Merge

Pin Fixture Hashes Before the Agent Merge

Comments
7 min read
Don't Ship Retries In the Bugfix Patch

Don't Ship Retries In the Bugfix Patch

Comments
6 min read
Survivor Mutants Block the Agent Merge

Survivor Mutants Block the Agent Merge

Comments
6 min read
Freeze Flakes Before Property-Checking Agent Patches

Freeze Flakes Before Property-Checking Agent Patches

Comments
7 min read
Revert Production. Agent Tests Must Fail.

Revert Production. Agent Tests Must Fail.

Comments 1
5 min read
Reject Agent Patches That Weaken the Oracle

Reject Agent Patches That Weaken the Oracle

Comments
7 min read
Don't Merge Agent Patches That Invent I/O

Don't Merge Agent Patches That Invent I/O

Comments
6 min read
Agent Patches Pass by Sharing Fixture State

Agent Patches Pass by Sharing Fixture State

Comments
7 min read
Pin the Clock Before You Merge an Agent Patch

Pin the Clock Before You Merge an Agent Patch

Comments
5 min read
Build for the 429: An Architecture for Unreliable Free LLM Tiers

Build for the 429: An Architecture for Unreliable Free LLM Tiers

Comments
4 min read
Property Checks and a Flake Freeze for Agent Patches

Property Checks and a Flake Freeze for Agent Patches

Comments
6 min read
Mutations Catch What Your Agent Patch Misses

Mutations Catch What Your Agent Patch Misses

Comments
3 min read
Catch Silent Behavior Drift in Agent Patches with Golden Snapshots

Catch Silent Behavior Drift in Agent Patches with Golden Snapshots

Comments
4 min read
Quarantine Agent Patches on a Free Ephemeral Server

Quarantine Agent Patches on a Free Ephemeral Server

Comments
4 min read
Fuzz the Patch, Not the Promise: A Hypothesis Harness for Agent-Generated Code

Fuzz the Patch, Not the Promise: A Hypothesis Harness for Agent-Generated Code

Comments
3 min read
A Zero-Token Triage for Agent-Patch Failures

A Zero-Token Triage for Agent-Patch Failures

Comments
4 min read
A Token-Efficient Gate for Agent Patches

A Token-Efficient Gate for Agent Patches

Comments
4 min read
A Testing Strategy for Agent Patches: Property Checks, Fixtures, and a Flaky Freeze

A Testing Strategy for Agent Patches: Property Checks, Fixtures, and a Flaky Freeze

Comments
4 min read
A Nightly Model-Regression Harness on Free Models and a Free Server

A Nightly Model-Regression Harness on Free Models and a Free Server

Comments
4 min read
A Response Cache for Token-Limited LLM APIs (Before You Burn the Free Grant)

A Response Cache for Token-Limited LLM APIs (Before You Burn the Free Grant)

Comments
4 min read
A Disposable Evaluation Loop for Any Model Release Before You Trust It in CI

A Disposable Evaluation Loop for Any Model Release Before You Trust It in CI

Comments
4 min read
A Reusable Smoke-Test Harness for Newly Released Open Models (Before You Bet a Project on One)

A Reusable Smoke-Test Harness for Newly Released Open Models (Before You Bet a Project on One)

Comments
6 min read
A Local AI Review Pass for Python Diffs Before You Open the PR

A Local AI Review Pass for Python Diffs Before You Open the PR

Comments
6 min read
A Pre-Push Gate for Python Diffs That Fails Loud, Not Often

A Pre-Push Gate for Python Diffs That Fails Loud, Not Often

Comments
5 min read
loading...