--- title: "Sentry Prompt Optimizer: review, setup & limitations · Undominated.ai" canonical: https://undominated.ai/skills/getsentry-prompt-optimizer/ description: "Starts from a written contract before rewriting a prompt, and names the model family as a choice rather than a default." --- # Sentry Prompt Optimizer: review, setup & limitations · Undominated.ai > Starts from a written contract before rewriting a prompt, and names the model family as a choice rather than a default. [← Explore all skills](/skills/) AI APPLICATION DEVELOPMENT / Sentry # Sentry Prompt Optimizer Starts from a written contract before rewriting a prompt, and names the model family as a choice rather than a default. [See setup guidance ↓](#setup)[Original source ↗](https://github.com/getsentry/skills/tree/d18b7aa8ba878354e5c348310230e652f7690f9c/skills/prompt-optimizer) SOURCE REVIEW Reviewed 2026-10-15 Evidence 3 linked sources Publisher Sentry Licence [Apache-2.0 ↗](https://github.com/getsentry/skills/blob/d18b7aa8ba878354e5c348310230e652f7690f9c/LICENSE) Revision d18b7aa8ba87 [Read what was—and wasn’t—checked ↓](#review) “Creates, optimizes, and iteratively refines agent prompts, system prompts, developer prompts, and reusable prompt templates. Use when asked to improve a prompt, optimize a system prompt, rewrite an agent prompt, tune prompt wording, make a prompt more reliable, port prompts between OpenAI, Claude, or Gemini, or build prompt evals.” [Sentry · upstream description ↗](https://github.com/getsentry/skills/blob/d18b7aa8ba878354e5c348310230e652f7690f9c/skills/prompt-optimizer/SKILL.md) Our analysis follows below. 01 / THE REASONING ## Why this made the selection. - Optimization starts from a written contract and an eval slice, not from a rewrite. - Different reference files are loaded for a new prompt versus an existing one. ### A good fit for - Rewriting a system or developer prompt against a named model family - Porting a prompt when the family is unknown, with a portable base and short adapter notes - Improve a prompt against a concrete failure set while preserving a holdout check. - Inventory external prompt context by exact path and distinguish loaded files from references and excluded material. ### Weigh up before choosing - The failure-mode list includes editing before the eval target is defined. - Durable policy is not supposed to live in the user payload. - Snapshot changes require evals to be rerun; the skill does not treat a previous score as still valid. - This review did not execute the skill. - Results on a small prompt-development set do not establish general reliability; model-specific advice can drift, and failures may require system changes rather than wording. 02 / THE REVIEW RECORD ## What we actually inspected. Source review has boundaries. A clear record is more useful than a “safe” badge. ### Material inspected - skills/prompt-optimizer/SKILL.md - skills/prompt-optimizer/references/core-patterns.md - LICENSE - vercel-labs/skills@48dc9e8eb8aa19040e91cb90ba0b642c9cbad57e:README.md ### Our findings - The pinned skill records task type, model snapshot, prompt surface, and success criteria before editing. - A new prompt loads core patterns, model-family notes, and transformed examples; an existing prompt loads the meta-optimization loop first. - Optimization baselines the current prompt on the same eval slice, clusters failures, and writes concrete edit criticisms. - The returned package is Target, Success Criteria, and External Context, with repo context cited by exact path. - The workflow inventories loaded versus merely referenced context, clusters failures, compares candidate changes and checks holdout behavior instead of optimizing by intuition alone. - External-context inventory uses repository-relative paths, distinguishes loaded, referenced, and out-of-scope files, and asks for only the excerpts required by the prompt or evaluation case. ### Not established by this review - The skill was not executed inside a host agent. - Not executed: it was not installed or run. The review applies to the material and revision named here. A newer upstream release can change its behavior. 03 / PUT IT TO WORK ## Add a skill to your workflow. [Upstream setup instructions ↗](https://github.com/vercel-labs/skills/blob/48dc9e8eb8aa19040e91cb90ba0b642c9cbad57e/README.md) DOCUMENTED COMMAND npx skills add getsentry/skills/skills/prompt-optimizer --skill prompt-optimizer Copy command ↗ Run in your terminal to choose your AI tools, project or global scope, and installation method. This short command follows the current upstream version; the review covers the revision shown above. Optional: install the reviewed revision Keeps the original source and installer version where recorded, while leaving your AI tools and installation scope for you to choose. npx skills add https://github.com/getsentry/skills/tree/d18b7aa8ba878354e5c348310230e652f7690f9c/skills/prompt-optimizer --skill prompt-optimizer Copy reviewed-revision command ↗ - Install the complete skill directory with its references, SPEC, and SOURCES; a copied SKILL.md alone is incomplete. - Choose your AI tools and project or global scope in the installer, then review its permissions before using the skill. ### Before you start - Node.js/npm for this installer - An agent that supports the Agent Skills format - The prompts under review and, for eval-backed work, a way to run them ### Compatibility Agent Skills-compatible coding agents · OpenAI, Claude, and Gemini model families for porting guidance ### Local Markdown instructions - Read the prompt and named context files such as AGENTS.md - No model API call is mandated by the file; evals are described as a step the agent runs when it can - Reads prompts and examples and may call a model API; redact sensitive cases and account for evaluation cost. ### Cost model Apache-2.0 licence; running evals against a model API needs that provider's account. 04 / FOLLOW THE EVIDENCE ## The source trail. Our notes are separate from the original resource. Check upstream before adopting a new version. - [Skill definition ↗](https://github.com/getsentry/skills/blob/d18b7aa8ba878354e5c348310230e652f7690f9c/skills/prompt-optimizer/SKILL.md) Checked 2026-10-14 https://github.com/getsentry/skills/blob/d18b7aa8ba878354e5c348310230e652f7690f9c/skills/prompt-optimizer/SKILL.md Supports: Workflow and stated scope - [Repository licence ↗](https://github.com/getsentry/skills/blob/d18b7aa8ba878354e5c348310230e652f7690f9c/LICENSE) Checked 2026-10-14 https://github.com/getsentry/skills/blob/d18b7aa8ba878354e5c348310230e652f7690f9c/LICENSE Supports: Redistribution terms - [Pinned continuation review W03 ↗](https://raw.githubusercontent.com/getsentry/skills/d18b7aa8ba878354e5c348310230e652f7690f9c/skills/prompt-optimizer/SKILL.md) Checked 2026-10-15 https://raw.githubusercontent.com/getsentry/skills/d18b7aa8ba878354e5c348310230e652f7690f9c/skills/prompt-optimizer/SKILL.md Supports: review.findings, bestFor KEEP COMPARING ## Other approaches to consider. Related by category or shared topics. These are alternatives to inspect, not a measured quality order. [### Agent Platform GenAI Inference ↗ Calls Gemini and third-party OpenMaaS models on Google Cloud Agent Platform, asking for model, project and region instead of defaulting.](/skills/google-agent-platform-inference/)[### Amazon Bedrock ↗ Bedrock model invocation, Knowledge Bases, Guardrails and AgentCore guidance, with quota, prompt-caching and cost-tracking workflows.](/skills/aws-amazon-bedrock/)[### Anthropic Claude API ↗ Routes Claude API work to language-specific SDK guidance and distinguishes self-hosted tool loops from managed agent execution.](/skills/anthropics-claude-api/) [AI Tools ↗](/tools/)[Skills ↗](/skills/)[Agents ↗](/agents/)[MCP Servers ↗](/mcp-servers/)[Workflows ↗](/workflows/) ## Continue your investigation - [Choose a toolkit](/tools/) - [Inspect execution hosts](/agents/) - [Check tool access](/mcp-servers/) - [Check project dependencies](/check-repo/)