CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

consulting-an-oracle

Use when stuck after multiple debug attempts and want to escalate to a stronger one-shot model (GPT-5 Pro, Opus, Gemini Pro) — packages a self-contained "oracle prompt" with Ruby/Rails project briefing, verbatim error, what-was-tried, constraints, and just-enough attached files. Triggers include "ask the oracle", "write a letter to GPT-5", "I'm stuck, draft a prompt for another model", "/tmp/letter.md".

Quality

83%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

63%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is well-structured and mostly lean, with genuinely concrete per-step guidance and a strong Ruby/Rails-specific suspect list. Its two real defects are a dangling reference to a nonexistent `template.md` (which both blocks the core output's layout and breaks progressive disclosure) and the absence of any verification step for the stated file/line caps and redaction pass.

Suggestions

Add the actual `template.md` to the bundle (e.g. references/template.md) containing the oracle-prompt section layout, or inline the section list in Step 6 — currently "Use the structure in `template.md`" points to a file that does not exist.

Close the validation gap: add a check after Step 6 (e.g. confirm total attached lines ≤2000 and re-grep the written file for the secret patterns before reporting) so the stated caps and redaction step are verified, not just declared.

Trim the two long rationale paragraphs (the model-orchestration decoupling rationale in Overview and the leak-recovery rationale in Step 5) to a single line each, and consider dropping or shortening the Quick Reference table that duplicates The Process.

DimensionReasoningScore

Conciseness

The body is mostly lean, imperative, and high-signal (concrete bash block, hard caps "≤8 files, ≤2000 total lines", terse anti-patterns), but the long rationale paragraphs — "keeping prompt-construction decoupled from model-orchestration keeps the skill model-agnostic (works with any oracle the user picks — Codex, Claude, Gemini, paste-into-chat) ..." and the swap-file/backup lecture in Step 5 — plus the Quick Reference table that duplicates The Process could be trimmed, matching the 4 anchor (efficient with minor over-explanation).

4 / 5

Actionability

Most steps are concrete (copy-paste bash for project detection, exact output path and invocation template, redaction regex), but Step 6 — "Use the structure in `template.md`" — defers the core deliverable's structure to a file that does not exist anywhere in the bundle, leaving the actual oracle-prompt layout missing; that is a missing key detail, matching the 3 anchor rather than 4's "minor gaps".

3 / 5

Workflow Clarity

Seven clearly sequenced steps with an explicit ordering constraint ("This step must run before Step 6, not after"), fallback paths, and a recap table match the 4 anchor (clear sequence, most checkpoints present); it is not 5 because there is no verification step — the caps (≤8 files / ≤2000 lines) and redaction are stated but never checked against the written file.

4 / 5

Progressive Disclosure

In-body structure is good (Overview → When to Use → Process → Anti-patterns), but the skill's only external reference — "See `template.md` for the exact section layout", referenced twice — points to a file that does not exist in the bundle (no references/, scripts/, or assets/ directories), so the deferred detail is unreachable; per the rubric's guidance to score against the actual bundle structure, this is a 3, not 4's "minor organization gaps".

3 / 5

Total

14

/

20

Passed

Description

95%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong description: it explicitly states what the skill produces, when to use it, and quotes natural trigger phrases a user would actually say, all scoped to a distinct niche. The only flaw is a minor voice slip ("want to escalate") into second person, which the rubric penalizes on specificity.

DimensionReasoningScore

Specificity

"packages a self-contained 'oracle prompt' with Ruby/Rails project briefing, verbatim error, what-was-tried, constraints, and just-enough attached files" lists multiple concrete actions and would match the 5 anchor, but the trigger clause "Use when stuck after multiple debug attempts and want to escalate" slips into second-person voice ("want"), which the rubric penalizes by reducing specificity by 1.

4 / 5

Completeness

Both what ("packages a self-contained 'oracle prompt' with ... verbatim error, what-was-tried, constraints, and just-enough attached files") and when ("Use when stuck after multiple debug attempts") are explicitly and concretely answered, matching the 5 anchor; it is not 4 because the 'when' is fully explicit with concrete trigger phrases rather than vague.

5 / 5

Trigger Term Quality

Quoted triggers "ask the oracle", "write a letter to GPT-5", "I'm stuck, draft a prompt for another model", and "/tmp/letter.md" comprehensively cover natural phrasings users would actually say, including a file-path trigger, matching the 5 anchor (comprehensive natural terms including synonyms and file extensions).

5 / 5

Distinctiveness Conflict Risk

Clear niche — escalating stalled Ruby/Rails debugging to a stronger one-shot model via a packaged prompt — with distinct triggers ("oracle", "letter to GPT-5") that no generic debugging or handoff skill would claim; minimal conflict risk per the 5 anchor.

5 / 5

Total

19

/

20

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository

lucianghinda/superpowers-ruby

Path

skills/consulting-an-oracle/SKILL.md

Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.