tech-specs

tech-specs is a skill for Claude Code, Codex from griddynamics/rosetta. It costs 23 tokens per session (1,110 once invoked), scanned A, original, Apache-2.0.

A skill for turning requirements into precise technical specifications, including the intended architecture, interfaces, contracts, test data, and checks for correctness.

In plain words
What is it for?
It is for documenting architectures, defining API contracts and data models, and producing specifications that can be tested or verified.
Why use it?
It removes ambiguity before implementation by describing what the finished system must do. It separates the target design from the step-by-step implementation plan.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/griddynamics/rosetta/tech-specs
Any agent
npx skills add griddynamics/rosetta --skill tech-specs
Clone the repo
git clone --depth 1 https://github.com/griddynamics/rosetta

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for tech-specs

README.md
[![agentmods](https://agentmods.dev/badge/skills/griddynamics/rosetta/tech-specs.svg)](https://agentmods.dev/skills/griddynamics/rosetta/tech-specs)
Your own site
<a href="https://agentmods.dev/skills/griddynamics/rosetta/tech-specs"><img src="https://agentmods.dev/badge/skills/griddynamics/rosetta/tech-specs.svg" alt="Measured on agentmods" height="20"></a>
Per session 23 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,110 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00023 $0.01110
Opus 5 $0.00012 $0.00555
Sonnet 5 $0.00005 $0.00222
Haiku 4.5 $0.00002 $0.00111

Measured yesterday against content hash 5799836a75fa, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

tech-specs scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

instructions/r2/core/skills/tech-specs/SKILL.md · 152 lines

How it starts

The opening of the file, as written. The whole thing — 152 lines — stays where its author put it; the contents beside it link to each section on GitHub.

<tech_specs>

Senior tech lead defining precise, testable technical specifications writing them compressed, concise, using terms always

<when_to_use_skill> Use when requirements need translation into specs, architecture needs documentation, or API contracts and data models need definition. Paired with planning skill: specs define WHAT (target state), plan defines HOW. Result defines complete target state with interfaces, contracts, test data, and verifiable criteria. </when_to_use_skill>

<core_concepts>

  • All Rosetta prep steps MUST be FULLY completed, load-context skill loaded and fully executed
  • Discovery MUST be completed before writing specs
  • MCPs and external sources MUST be used to acquire context (DeepWiki, Context7, Web Search)

Tech specs define target state; plan defines steps to reach it. Split with companion planning skill: specs own WHAT, plan owns HOW. Do NOT repeat across both. Keep consistent. When one changes, verify the other.

Tech Spec Flow:

  1. Write TOC first
  2. Write section by section (do NOT write entire document at once)
  3. Verify integrity as separate step (do not combine with writing)
  4. Insert TLDR at the beginning (up to 10 lines)

Spec sections (adapt per request):

  1. Overview & Scope & TLDR
  2. Non-Functional Requirements and Architecture Significant Requirements
  3. Architecture & Component Design
  4. API Contracts
  5. Data Models & Schemas
  6. Error Handling Strategy
  7. Testing Strategy with Test Cases
  8. Security Considerations
  9. Dependencies
  10. Assumptions
  11. Tech Summary: files and services affected

</core_concepts>

<request_size_scaling>

Scale per request size classification:

SMALL MEDIUM LARGE
Output message, no files concise specs file, light and short full specs document
Sections overview + affected areas core sections all sections
Detail concise, signatures only signatures + contracts full specs
Length up to 100 lines 100-200 lines 200-500 lines
Diagrams none key interfaces sequence + component
Security skip unless critical threat summary full STRIDE

</request_size_scaling>

<spec_rules>

  1. Adapt to request size per scaling table
  2. Audience: senior engineers; do not explain obvious
  3. Compact, dense, complete
  4. Interfaces, signatures, contracts, API specs, endpoints
  5. Sequence diagram when 4+ actors involved
  6. Domain-specific patterns only; mention standards and best practices without explaining them
  7. Shorter is better
  8. Logically structured per project context
  9. Detail down to interfaces/classes/methods (signatures only, no implementations)
  10. Accuracy over speed
  11. Code snippets max 3 lines, only when critical

</spec_rules>

<design_principles>

Specs MUST follow: SRP, SOLID, KISS, DRY, YAGNI, MECE. Reference these when defining component boundaries, interfaces, and responsibilities. Do not explain the principles — apply them.

</design_principles>

<security_considerations applies="security-critical features: auth, payments, PII, FedRAMP">

  • Threat model (STRIDE)
  • Attack vectors and mitigations
  • Compliance requirements (GDPR, SOC2, etc.)
  • Security testing requirements

</security_considerations>

<test_data_considerations>

  • Happy path examples (3-5 cases)
  • Edge cases (boundary values)
  • Error cases (malformed input)
  • Security test cases (injection, tampering)
  • Performance test parameters (load, concurrency)

</test_data_considerations>

<validation_checklist>

  • TLDR present and accurate (up to 10 lines)
  • All sections internally consistent
  • Interfaces defined down to method signatures
  • Sequence diagrams present when 4+ actors
  • Security section present for security-critical features
  • Test data covers happy path, edge, error, security cases
  • No duplication with companion plan
  • Specs match companion plan
  • Scaled appropriately to request size
  • Engineers can implement without further clarification

</validation_checklist>

<best_practices>

  • Start from approved discovery and requirements
  • Use terms, abbreviations, diagrams over prose
  • Wrap specs output with <CRITICAL ATTRIBUTION="DO NOT COMPACT/OPTIMIZE/SUMMARIZE/REPHRASE, PASS AS-IS">...</CRITICAL>

Read the full file on GitHub · 152 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 152 lines · 23 tokens per session scan A 5799836a75fa

Subscribe to this mod's changes

tech-specs is a skill published in the GitHub repository griddynamics/rosetta (342 stars, last pushed yesterday), licensed Apache-2.0. It adds 23 tokens to every session and 1,110 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.