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.
npx agentmods add skills/griddynamics/rosetta/init-workspace-rulesnpx skills add griddynamics/rosetta --skill init-workspace-rulesgit clone --depth 1 https://github.com/griddynamics/rosettaWhat 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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00010 | $0.00981 |
| Opus 5 | $0.00005 | $0.00491 |
| Sonnet 5 | $0.00002 | $0.00196 |
| Haiku 4.5 | $0.00001 | $0.00098 |
Grade A, and why
init-workspace-rules 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 2d ago.
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.
How it starts
The opening of the file, as written. The whole thing — 102 lines — stays where its author put it; the contents beside it link to each section on GitHub.
<init_workspace_rules>
<when_to_use_skill> Local copies of Rosetta instructions enable AI agents to load rules without Rosetta access and stay current via periodic version checks. Creates full local files for all Rosetta content adapted to detected IDE/CodingAgent format. Validation: all Rosetta content exists as local files, root entry point triggers full prep chain. </when_to_use_skill>
<core_concepts>
- All Rosetta prep steps MUST be FULLY completed, load-context skill loaded and fully executed
- Rules consumed by AI agents, not humans
- Full-copy mode — copies complete file content from Rosetta to local workspace
- Adapt — copy content AS-IS; adapt ONLY IDE format: extension, frontmatter, directory. Never rewrite instruction content.
- Exclusion set —
init-workspace-*skills/workflows,templates/shell-schemas/*,configure/*,rules/bootstrap.mdMUST NOT BE copied - Bundled ACQUIRE — when ACQUIRE returns multiple
<rosetta:file>sections, strip tags, merge into one file with one frontmatter - state.mode —
initcreates all files;upgradefills gaps only, never overwrites human-customized files - Make sure that you follow original activation conditions, MUST never make all rules to be ALWAYS activated/loaded (overflows context)
</core_concepts>
Internal knowledge about IDE/agent configuration is obsolete — LIST and ACQUIRE from KB.
Step 1: Identify Environment
- LIST
configureIN KB with XML format (to understand supported IDE/CodingAgents) - Detect current environment, preselect IDE/CodingAgent
- MUST ask user to confirm selection and provide multi-choose
- ACQUIRE FROM KB
- If multiple selected, use common standards to reduce copies
Step 2: Read Workspace Context
- Read TECHSTACK.md and relevant project docs
Step 3: Discover Full Rosetta Content (subagent)
- LIST
allFROM KB with format=flat, save to FEATURE TEMP folder aslist-all-output.md - Parse into content-type groups (rules, skills, agents, workflows, commands)
- Apply exclusion set
- Report: total count, per-type count, excluded count
Step 4: MUST Install Root Entry Point and Bootstrap Rules
- ACQUIRE
rules/local-files-mode.mdFROM KB — install as root entry point per IDE configure spec - Embed Rosetta version marker (e.g., "R2.0") in core root file for staleness detection
- Apply IDE-specific frontmatter format from configure file
- ACQUIRE each
rules/bootstrap-*.mdFROM KB — install as individual rule files per IDE configure spec
Step 5: MUST Generate All Content Files
For each content type from filtered list (non-bootstrap rules, skills, agents, workflows, commands):
- Map ResourcePaths to local file paths using configure file rules
- If state.mode=upgrade: skip existing human-customized files
- ACQUIRE each resource FROM KB
- Write to local path with IDE-specific format adaptation
- Preserve skill subdirectory structures (assets/, references/, scripts/)
- If multiple IDEs: write shared content to common location where possible
Step 6: Verify and Report (HITL)
- Count files per type, compare against expected from filtered list minus exclusions
- Verify: no absolute paths in generated files
- Verify: root entry point file contains version marker
- Verify: bundled ACQUIRE content merged correctly (no
<rosetta:file>tags, single frontmatter per file) - If state.mode=upgrade: report diff summary (added, skipped with reason)
- MUST get explicit user confirmation before closing
<validation_checklist>
- Agent with no prior context can bootstrap from generated files using only local filesystem
- Every content type from LIST output has corresponding local files (none silently dropped)
</validation_checklist>
- ACQUIRE/SEARCH/LIST commands inside instruction content are local-files-mode aliases — do NOT remove or replace them
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.
- 2d ago First seen · 102 lines · 10 tokens per session scan A 03b3ee8f7913
init-workspace-rules is a skill published in the GitHub repository griddynamics/rosetta (341 stars, last pushed 4d ago), licensed Apache-2.0. It adds 10 tokens to every session and 981 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 2026-08-30.
Other skills, from other repositories
authoring-skills-and-rules
Author Skills (SKILL.md) and rules/instructions across Claude Code, Cursor, OpenCode, Codex, and GitHub Copilot. Covers frontmatter, file layout, activation model, cross-platform mirroring, and the universal craft: progressive disclosure, trigger-rich descriptions (context pointers), leading words, token budget…
memorix-orchestrate
Use when a main agent needs Memorix to coordinate explicit subagent work through tasks, handoffs, messages, file locks, or the orchestrate CLI.
skill-builder
Meta-skill for creating new skills from natural language. Trigger phrases: "new skill", "create a skill", "build a skill", "make a skill", "generate a skill", "author a skill", "skill builder".
agents-md-author
Bootstrap or update project AI-agent instruction files (AGENTS.md for Codex, CLAUDE.md for Claude Code, or both) from the current project's structure. Triggers on: write AGENTS.md, write CLAUDE.md, generate AGENTS.md, agent instructions, project conventions for AI, bootstrap agent docs, set up project memory for…
skill-creator
Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends Claude's capabilities with specialized knowledge, workflows, or tool integrations.
framework-builder
Build reusable agent-readable frameworks from any domain expertise. Use when someone says "systematize this," "turn this into a repeatable process," "build a framework for," "create a skill for," "I keep doing the same thing," or "make this predictable." Produces a complete five-layer framework in markdown that agents…