Skip to content
Agentmart
Submit a tool

agentic-core-tool-first

Establishes tool-first architecture where the model plans and validates while tools perform concrete operations under strict permissions.

Skill by rupeshitpatekar

IndexedMIT★ 1
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/rupeshitpatekar-agentic-core-tool-first

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Establishes tool-first architecture where the model plans and validates while tools perform concrete operations under strict permissions.

Reviews

Write a review

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/agentic-core-tool-first

What's inside

Skill name
agentic-core-tool-first
Description
Establishes tool-first architecture where the model plans and validates while tools perform concrete operations under strict permissions.
License
MIT
Instructions
40 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Agentic Core: Tools Do the Real Work

Use this skill to keep operational actions in tools and keep model behavior supervisory.

## Principles

1. Tool calls are explicit, typed actions.
2. Tool permissions are least-privilege and auditable.
3. Unsafe operations require additional policy checks.

## Tool contract template

- input schema
- preconditions
- side effects
- output schema
- failure semantics

## Execution rules

1. Validate tool args before execution.
2. Log decision and correlation ID.
3. Execute tool and capture structured output.
4. Surface errors explicitly; do not mask.

## Required output format

1. **Tool Inventory**
2. **Permission Matrix**
3. **Execution Policy**
4. **Error and Retry Policy**
5. **Abuse/Misuse Controls**

## Abuse controls

- denylist for destructive operations unless explicitly approved.
- bounded retries for non-idempotent operations.
- redaction/secret handling for tool outputs.

From rupeshitpatekar/claude-code-leak-learnings, MIT.

More from rupeshitpatekar/claude-code-leak-learnings

Signals

Stars
1
Forks
0
Open issues
0
Last commit (this folder)
Apr 4, 2026 (6 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
Repository
github.com/rupeshitpatekar/claude-code-leak-learnings
First seen
Oct 24, 2026
Last checked
Oct 24, 2026
Source
GitHubgithub-graphql

More from rupeshitpatekar

Cite this page

agentic-core-tool-first skill. Agent Mart. https://agentmart.dev/skills/rupeshitpatekar-agentic-core-tool-first (last updated 2026-10-24)