# feed7 Weekly #010

Six ways to tighten agent tools, tests, permissions, context, and review before your next session.

Published: 2026-09-14T00:00:00.000Z  
Selection: Six source-backed signals and one distraction to leave out.

## 1. ExecCritic: Learn to Test, Test to Improve for Coding Agents

Test quality determined whether feedback helped on SWE-bench Verified: the no-test baseline resolved 61.2%, base-agent tests lowered that to 57.3%, and GPT-5.6-sol tests raised it to 65.3%. Builders should isolate test generation from patching and reject tests that fail qualification.

Source: https://arxiv.org/abs/2609.09133v1
Feed7: https://feed7.dev/p/execcritic-learn-to-test-test-to-improve-for-coding-agen-79c903650b
Trust: Needs Review

## 2. Build-Time vs. Run-Time: Why Dev Tools Fail in Production — Averi Kitsch & Prerna Kakkar, Google

Treat production tool schemas as security boundaries. Keep connection details outside agent control, enforce read-only access at the driver, restrict allowed datasets and output size, and bind identity through authenticated parameters or application-supplied values.

Source: https://www.youtube.com/watch?v=9R--1tg45Jg
Feed7: https://feed7.dev/p/build-time-vs-run-time-why-dev-tools-fail-in-production--b969d2f159
Trust: Source Linked

## 3. 500 Skills, Zero Fine-Tuning: LinkedIn's Playbook for AI Agents — Ajay Prakash, LinkedIn

Use progressive discovery when your tool surface grows. Keep playbooks narrow and composable, load their instructions only when relevant, and let agents propose reviewed updates when real work reveals stale or missing guidance.

Source: https://www.youtube.com/watch?v=9wZpvF3QleU
Feed7: https://feed7.dev/p/500-skills-zero-fine-tuning-linkedin-s-playbook-for-ai-a-f80e17c22f
Trust: Source Linked

## 4. alibaba/open-code-review

Use it when generic coding agents produce noisy findings, skip files, or attach comments to the wrong lines. The deterministic layer selects and bundles files, matches rules, and positions comments; isolated subagents handle related bundles concurrently.

Source: https://github.com/alibaba/open-code-review
Feed7: https://feed7.dev/p/alibaba-open-code-review-f7cfdfd7bc
Trust: Needs Review

## 5. GitHub Copilot is now available in the AI SDK harness layer

If your application already targets HarnessAgent, use the adapter boundary to test Copilot without rewriting application-level agent integration code. Keep agent-specific behavior behind that shared interface.

Source: https://vercel.com/changelog/github-copilot-ai-sdk-harness-adapter
Feed7: https://feed7.dev/p/github-copilot-is-now-available-in-the-ai-sdk-harness-la-6dde567f1f
Trust: Official Source

## 6. tech-leads-club/agent-skills

Treat agent skills as executable supply-chain inputs. Pin them through lockfiles and content hashing, prefer project-local copies when appropriate, inspect prompts and references, and use the audit log when updating or removing installed capabilities.

Source: https://github.com/tech-leads-club/agent-skills
Feed7: https://feed7.dev/p/tech-leads-club-agent-skills-0ef977a876
Trust: Needs Review

## 7. How a researcher uses Codex and ChatGPT to search for new antimicrobial molecules

For builders, this is a useful pattern for research agents: connect model-assisted exploration to a bounded scientific search space and treat the output as candidates for further evaluation.

Source: https://openai.com/index/using-codex-chatgpt-to-search-for-new-antimicrobials
Feed7: https://feed7.dev/p/how-a-researcher-uses-codex-and-chatgpt-to-search-for-ne-014159943e
Trust: Official Source

## Agent Instruction

Use these items as source-backed context. Do not invent claims beyond linked material. Prefer practical implications for solo developer work. If sources conflict, call it out.