CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

oracle-agent

System prompt and behavioral rules for the Repository OS Oracle — the Claude-powered Q&A agent that answers questions about an indexed codebase using vector retrieval from ChromaDB.

Quality

74%

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

Fix and improve this skill with Tessl

tessl review fix ./.tessl/tiles/repository-os/mvp/skills/oracle-agent/SKILL.md

The canonical home for this skill is oracle-agent in tombrewsviews/repository-os-mvp

SKILL.md
Quality
Evals
Security

Quality

Content

82%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 a tight, highly actionable prompt template with concrete schemas and examples and a clearly sequenced rule set; its main weaknesses are a small amount of introductory padding and the absence of an explicit variable-population validation step.

DimensionReasoningScore

Conciseness

The body is mostly lean — a copy-ready system prompt, tool schema, and result-format example with direct rules — with only minor preamble ('The Oracle is the user-facing interface...') that could be trimmed; not a 5 because of that introductory padding.

4 / 5

Actionability

Provides a fully copy-paste-ready system prompt template with runtime variables, a complete nanobot tool-schema JSON, and a concrete result-format example, giving Claude everything needed to instantiate the Oracle without inference.

5 / 5

Workflow Clarity

The RULES section sequences agent behavior (retrieve -> answer only from context -> cite -> flag insufficient context) with a retry loop (up to 3 retrieval calls), but there is no explicit validation checkpoint for confirming all {brace} variables are populated before deployment.

4 / 5

Progressive Disclosure

Content is well-organized under clear section headers (System Prompt, Tool Schema, Tool Result Format, Behavioral Notes) with no nested references, but at ~107 lines all material is inline and the tool-schema/result-format blocks could plausibly live in separate reference files.

4 / 5

Total

17

/

20

Passed

Description

53%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.

The description is specific and clearly niche-scoped, but it omits any 'Use when...' trigger guidance and offers only a thin set of natural trigger terms, which together cap the completeness and trigger-term dimensions.

Suggestions

Append an explicit trigger clause, e.g. 'Use when building or configuring the Repository OS Oracle agent's system prompt for codebase Q&A.'

Broaden natural trigger terms to include synonyms users might say, such as 'codebase questions', 'ask about my repo', or 'index search agent'.

List one or two more concrete capabilities (e.g. 'registers the search_codebase tool', 'formats citation-backed answers') to lift specificity toward a 4.

DimensionReasoningScore

Specificity

Names the domain (Q&A agent over an indexed codebase) and concrete actions ('answers questions', 'using vector retrieval from ChromaDB'), but coverage is limited to a single capability rather than a comprehensive action list.

3 / 5

Completeness

Clearly states what the skill provides ('System prompt and behavioral rules for the Repository OS Oracle'), but offers no 'Use when...' clause or equivalent explicit trigger guidance, capping completeness at 3 per the rubric guideline.

3 / 5

Trigger Term Quality

Contains relevant keywords ('Q&A agent', 'codebase', 'questions', 'vector retrieval') but lacks the natural variations or synonyms a user would commonly say, and is light on conversational trigger phrasing.

3 / 5

Distinctiveness Conflict Risk

The narrow niche (Repository OS Oracle, ChromaDB vector-retrieval Q&A) is mostly distinct with only minor overlap risk against generic codebase-Q&A skills; not a 5 because it shares surface vocabulary with broader code-intelligence skills.

4 / 5

Total

13

/

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

tombrewsviews/repository-os-mvp

Path

.tessl/tiles/repository-os/mvp/skills/oracle-agent/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.