Skip to content
Skills Library
AI & agents · Planning

deterministic-judgment

/deterministic-judgment

Use when adding, reviewing, or choosing deterministic logic versus model-led reasoning, especially classifiers, validators, rule engines, source selection, legal/tax/compliance workflows, LLM fallbacks, or any change justified by determinism, repeatability, or token savings.

Repository
app_files ↗
Licence
AGPL-3.0
Last repository change
27.10.2026

Assessment and security

90/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content25 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
1
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 27.10.2026, without references to retired models or APIs.

Licence

AGPL-3.0, copyleft: przeróbki udostępniane innym muszą mieć tę samą licencję.

What it does and when to use it

Use when adding, reviewing, or choosing deterministic logic versus model-led reasoning, especially classifiers, validators, rule engines, source selection, legal/tax/compliance workflows, LLM fallbacks, or any change justified by determinism, repeatability, or token savings.

Original author description

Use when adding, reviewing, or choosing deterministic logic versus model-led reasoning, especially classifiers, validators, rule engines, source selection, legal/tax/compliance workflows, LLM fallbacks, or any change justified by determinism, repeatability, or token savings.

How it works

  1. Required Standard
  2. Decision Rule
  3. Implementation Requirements
Original SKILL.md headings · 58 lines, 382 words
SKILL.md preview
---
name: deterministic-judgment
description: Use when adding, reviewing, or choosing deterministic logic versus model-led reasoning, especially classifiers, validators, rule engines, source selection, legal/tax/compliance workflows, LLM fallbacks, or any change justified by determinism, repeatability, or token savings.
---

# Deterministic Judgment

Deterministic is not preferred by default. Use deterministic logic only when it
is demonstrably better for the specific job than model-led reasoning.

## Required Standard

Before adding or preserving deterministic logic, state the reason it should be
deterministic. A valid reason must identify at least one concrete advantage:

- mechanically verifiable correctness;
- lower observed error rate than model-led reasoning on representative cases;
- strict reproducibility required by an external contract;
- security, safety, or auditability that depends on fixed rules;
- performance or cost savings after quality is shown not to regress.

Invalid reasons:

- deterministic is better in theory;
- deterministic saves tokens;
- deterministic feels safer;
- deterministic makes the system look more engineered;
- the task has keywords that can be classified with simple string matching.

## Decision Rule

Use deterministic code for mechanical work: parsing stable formats, schema
validation, exact arithmetic, file packaging, cache keys, ID matching, simple
presence checks, and transformations with explicit contracts.

Use model-led reasoning for semantic judgment: legal or tax relevance, source
selection, topic taxonomy, research scope, document meaning, claim
interpretation, ambiguous intent, and anything where context changes the answer.

Hybrid designs are acceptable only when the boundary is clear: deterministic
code may collect evidence or enforce output shape, but it must not overrule
model judgment on semantic relevance unless tests show it is better.

## Implementation Requirements

- Include a short comment, docstring, or PR note explaining why any new
  deterministic rule is justified.
- Add representative tests for deterministic rules, including false-positive
  and false-negative cases when the rule gates behavior.
- For legal, tax, or compliance modules, do not let deterministic classifiers
  choose legal topics, governing framework, source domains, or research phasing.
- If deterministic output disagrees with model judgment on a semantic issue,
  prefer model judgment unless a documented benchmark or external requirement
  says otherwise.
- Remove deterministic logic that exists only because it is deterministic and
  has no demonstrated quality advantage.
- When reviewing an execution block, trace the rejected condition to the
  concrete failure it prevents. An existing rejection test establishes current
  behavior, not that the rejection is necessary. Check a valid case that should
  proceed as well as the failure case. External version labels and installation
  history alone do not establish missing runtime capabilities.

Versions

The database has no other skills named deterministic-judgment.

Database snapshot: 13.11.2026 · package version 34f5a05adb8f (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

No similar skills in the database.

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.