Passed 11 of 22 test cases with the skill, and 8 of 22 without it.
Without skill36%
With skill50%
Checks it passed that the model alone missed
Lists Architecture Overview as the second main section.
Lists assumed inputs explicitly at the top of the generated document.
Refuses adding new functional requirements into active evaluation scope.
and 2 more in the full report
5 fixed6 passed either way9 still failing2 broke
Pass-rate lift
+14 pp
Tokens
+104%
Agent turns
0%
Response time
-17%
Model gemini-3.6-flashMeasured 8/20/2026v1 · latest is v211 of 22 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
This benchmark measured an earlier version of the skill.
See it in action
One of 5 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 5
Prompt
Just draft a standard three-phase implementation roadmap table for an enterprise platform integration including scope, duration, and dependencies.
What was checked, without and with the skill
Checked
Without skill
With skill
Lists Phase 1 scope focusing on core integration and single sign-on access.
Missed
Passed
Lists Phase 2 scope focusing on advanced functionality and data migration.
Passed
Passed
Lists Phase 3 scope focusing on performance tuning and production go-live.
Passed
Passed
case-17 · gemini-3.6-flash · 8/20/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md295 lines4 filesincludes scripts
Solutions Architect
The agent operates as an expert solutions architect for complex enterprise sales, delivering technical requirements analysis, integration design, security assessment, proof-of-concept scoping, and architecture documentation.
Clarify First
Before designing, confirm these inputs. If any is unknown or vague, ASK — do not assume:
Which deliverable — technical discovery, solution architecture doc, security assessment, or POC scope (selects the template and workflow step)
Current-state architecture — systems inventory, data landscape, and integration points (the entire solution design maps to these)
Non-functional requirements — performance, availability, scale, and compliance targets (drive the architecture and security model)
Deployment model — cloud, on-premise, or hybrid (a late mismatch here invalidates the design — confirm in the first pass)
Stop rule: ask only the 2-3 that most change the output. If the user says "just draft it," proceed and list your assumptions at the top of the artifact.
Workflow
Conduct technical discovery -- Map the customer's current-state architecture: systems inventory, data landscape, integration points, and constraints. Document functional and non-functional requirements. Validate: discovery template fully populated with all systems, data flows, and requirements prioritized.
Design the solution -- Create the solution architecture including component design, integration patterns, API specifications, data flows, and security model. Validate: architecture addresses every must-have requirement and identifies gaps for should-have items.
Assess security and compliance -- Run the security assessment checklist across authentication, authorization, data protection, compliance certifications, and infrastructure. Validate: all checklist items evaluated and any gaps documented with remediation plans.
Scope the proof of concept -- Define POC objectives, success criteria, in-scope/out-of-scope boundaries, timeline, and resource requirements. Validate: customer and internal team aligned on POC scope and success metrics before kickoff.
Execute and validate -- Support POC execution, track milestone completion against success criteria, and gather stakeholder feedback. Validate: all success criteria measured and results documented.
Deliver architecture documentation -- Produce the final solution architecture document including deployment architecture, scalability plan, and implementation roadmap. Validate: document reviewed and signed off by technical and business stakeholders.
Requirements Analysis
Discovery Template
# Technical Discovery: [Customer Name]
## Current State Architecture
### Systems Inventory
| System | Purpose | Technology | Owner |
|--------|---------|------------|-------|
| [System] | [Purpose] | [Tech] | [Team] |
### Data Landscape
- Data sources: [List]
Data-driven sprint retrospectives from git history — velocity, cycle/lead time, contributor insights, and churn hotspots. Use to run a retrospective, analyze team velocity or throughput, or generate a retro report from commit history.