Find agent skills by outcome

docs/ai-instructions.md

generic skill

- Gemini 2.5 Pro / Flash (Google) — Outstanding multimodal (especially video/document understanding)

very large context windows

- Claude 4 Opus/Sonnet (Anthropic) — Exceptional long-form reasoning

writing quality

- Llama 4 / Llama 405B variants (Meta) — Best open-source frontier performance

excellent for self-hosting

- Tool use

function calling

- Grok 3 / Grok 4 (xAI) — Excellent reasoning

real-time knowledge via X

Current Best-in-Class Ranking (January 2026 – general guidance

always tailor to the process):

Productivity

- GPT-5 / o3 family (OpenAI) — Deepest reasoning on very complex structured tasks

best at following extremely long/complex instructions

Design

Cog came from Andreas's experience at Spotify and Ben's experience at Docker.

- Privacy

open behavior

If AI is viable

recommend which AI engines are best suited and why.

| Data quality | 3 | Mostly clean

some inconsistency |

When scoring

evaluate these dimensions (suggested weights for averaging: e.g.

Summarize the overall feasibility score (weighted average)

then issue your verdict with clear reasoning.

Interpretation: Clear patterns exist

but context accuracy is critical. Recommend hybrid approach with AI drafts + human review.

- Call out misconceptions

dependencies

- Track KPIs (e.g.

20% time savings

| 4 – High Feasibility | Clear logic

consistent data

| 5 – Excellent Feasibility | Predictable process

well-defined data

Productivity

| 3 – Moderate Feasibility | Certain tasks could be automated (e.g.

drafting

- What is your budget or resource availability for AI implementation (e.g.

time

- How would you measure the success of AI support (e.g.

time saved

After the interview

provide a structured assessment.

| 2 – Low Feasibility | Some structured elements exist

but goals or data are unclear. AI could assist with insights