# feed7 Weekly #002

7 AI signals worth feeding your agents this week.

Published: 2026-07-07T00:00:00.000Z  
Selection: Six source-backed signals and one distraction to leave out.

## 1. Gemini agent SDK adds shared context caching across sessions

Big static context (style guides, schemas, docs) belongs in the cache, not the prompt. Restructure bundles so stable material leads.

Source: https://developers.googleblog.com/gemini-agent-sdk-context-cache
Feed7: https://feed7.dev/p/gemini-context-caching
Trust: Official Source

## 2. OpenAI ships strict structured outputs for parallel tool use

Delete your hand-rolled tool-call validators. Turn on strict mode and move validation effort to eval coverage instead.

Source: https://openai.com/blog/structured-outputs-parallel-tools
Feed7: https://feed7.dev/p/openai-structured-tool-use
Trust: Official Source

## 3. Building eval sets that survive model swaps — AI Engineer workshop

Rewrite phrasing-based assertions as outcome assertions now — before your next model swap forces it.

Source: https://www.youtube.com/watch?v=ai-eng-evals-2026
Feed7: https://feed7.dev/p/evals-workshop-aie
Trust: Transcript Verified

## 4. ctxlint: a linter for agent context files

Context files rot silently. Run ctxlint in CI so your agent instructions stay as maintained as your code.

Source: https://github.com/ctxtools/ctxlint
Feed7: https://feed7.dev/p/ctxlint-repo
Trust: Repo Verified

## 5. Conductor: a Mac app that runs local agent fleets

Parallel agents stop stepping on each other. Worth adopting if you run more than one coding session a day.

Source: https://conductor.build
Feed7: https://feed7.dev/p/conductor-mac-app
Trust: Founder Tested

## 6. Vercel dashboard density — a layout taste reference

Decision: adopt hairline-border density for the Brain library instead of spaced-out cards. Maps to spacing tokens 16/24.

Source: https://vercel.com/design/dashboard-density
Feed7: https://feed7.dev/p/vercel-dashboard-density
Trust: Source Linked

## 7. Rolling out agents behind evals — an operator’s playbook

The staging pattern is reusable today; treat the win-rate numbers as unverified until the promised write-up lands.

Source: https://www.linkedin.com/posts/operator-evals-rollout
Feed7: https://feed7.dev/p/linkedin-eval-rollout
Trust: Needs Review

## 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.