Agent Change Reviewer — Development agent for Claude Code
Reviews a proposed change to a Sarvam agent (prompt diff, configure_agent payload, tools.py) for regressions, list-replacement mistakes, undefined variables/tools/states, language/TTS issues and safet.
How to install Agent Change Reviewer
Installs to ~/.claude/agents/atishayganesh-sarvam-kit-agent-change-reviewer.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/AtishayGanesh/sarvam-kit/HEAD/agents/agent-change-reviewer.md -o ~/.claude/agents/atishayganesh-sarvam-kit-agent-change-reviewer.md Restart Claude Code, or start a new session, for it to be picked up.
What Agent Change Reviewer does
name: agent-change-reviewer description: Reviews a proposed change to a Sarvam agent (prompt diff, configure_agent payload, tools.py) for regressions, list-replacement mistakes, undefined variables/tools/states, language/TTS issues and safety gaps before it is sent. Use before any configure_agent or upload_agent_code call. tools: Read, Grep, Glob, Bash
You are the last check before a platform mutation. Given the current config, the proposed payload/diff and the repo's skills, verify: eve
Alternatives in Development
- Voice Engineer — GAIA voice interaction specialist 1.5k ★
- Refactor Validator — Use this agent when a significant refactoring has been completed and you need to verify that the changes are c 358 ★
- Part 13: Memory Bridge (Give Coding Agents Your Brain) — When you spawn Codex or Claude Code to build something, they start blind 232 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Deal Editor
Drafts replacement contract language for a flagged clause. Produces ready-to-paste text in the contract's own
Options Risk Reviewer
Adversarially review a proposed option structure before any capital is committed. Re-derives the risk independ
Operational Readiness
Assesses the ongoing human cost of the proposed feature — support burden, moderation/ops staffing, maintenance
Conversation Analyst
Reads a slice of a redacted conversation corpus for one Sarvam agent and returns a failure taxonomy with count
Reverse Salient Agent
Surfaces reverse-salient findings (Engine 1 Act 1) as F.0 Decision Gates with persona-aware framing. Sibling t
Tabz Expert
Browser automation and terminal spawning expert with 82 MCP tools. Use when the user asks to 'automate the bro
Related Skills
Review Changes
Use when asked to review changes in this repository. Focus on bugs, regressions, API contract drift, retrieval
Security Audit Web App
Use when auditing a web app frontend/edge layer for common security mistakes — server vs client boundary leaks
Dsh Plugin Hooks
DeepSeek Harness plugin: Claude-Code-style lifecycle hooks for DSH. Configure shell commands that run automati