Frozen, deterministic decision (no LLM): maps project_license, dependency_license, linking, distributing, modified_dependency to one of ok_internal_use, must_offer_source, project_must_be_gpl, compatible, incompatible, review_required.
SkillScore68Maintained
Automatic
◆Legal & Contracts pack
Last verified 8/4/2026
What it changes
Passed 13 of 19 test cases with the skill, and 4 of 19 without it.
Without skill23%
With skill68%
Checks it passed that the model alone missed
Invokes assess_license with linking set to static
Passes linking as dynamic and distributing as True
Passes distributing as None when calling assess_license
and 13 more in the full report
10 fixed5 passed either way7 still failing0 broke
Pass-rate lift
+45 pp
Tokens
-36%
Agent turns
0%
Response time
-52%
Model gemini-3.6-flashMeasured 8/4/2026v115 of 22 cases passed with skill · 19 comparable
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 10 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 10
Prompt
We distribute a CLI utility under MIT license that imports an unmodified MIT-licensed utility package directly via package registry. Evaluate whether this configuration is compliant and return the verdict.
What was checked, without and with the skill
Checked
Without skill
With skill
Executes assess_license with MIT for both project_license and dependency_license
Missed
Passed
Returns the verdict output produced by the decision tree script
Missed
Passed
case-04 · gemini-3.6-flash · 8/4/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md30 lines6 filesincludes scripts
assess_license — skill (tempered by temper-skills)
You are an assistant.
The decision is frozen. Do not re-derive it from prose or your own judgment — the routing logic now lives in a deterministic decision tree (assess_license.assess_license, zero LLM calls, reviewed and version-controlled). Your job is the part the tree cannot do: turn the request into structured features, call the tree, and phrase its verdict.
How to answer
Extract these structured features from the request:
project_license
dependency_license
linking
distributing
modified_dependency
Call the decision tree and treat its result as authoritative (bundled at scripts/assess_license.py):
Relay verdict to the user. Do not override it. If a feature can't be extracted, pass it as None — the tree is built to fall through safely.
Gray zones to surface
The tree flags these as underdetermined — mention the caveat when the answer touches them:
(n2) network-use scope is fact-specific — confirm deployment model
(n6) LGPL static-linking obligations are satisfiable (provide object files) — counsel can clear it
Generated by temper-skills from the original skill · 2026-07-01T12:45:19Z · model: claude-opus-4-8 via temper-skills. The decision logic is now testable (temper-skills validate) and evolvable (temper-skills incremental) — regenerate this skill when the tree changes.