Memory Tool And Context Editing
Claude API · Anthropic
Foundation Model & API — Anthropic (Claude)
Evaluates Anthropic's Memory Tool & Context Editing across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Foundation Model & API eval coverage.
About Anthropic
Anthropic is an AI safety company and the maker of Claude. Its API exposes the Claude model family (Opus, Sonnet, Haiku) with tool use, prompt caching, extended thinking, batch processing, vision, the Files and Memory tools, and the Claude Agent SDK.
Sample tests· showing 3 of 9
| # | Input | Expected behavior | Check |
|---|---|---|---|
| 01 | Agent includes the memory tool (type=memory_20250908) in tools[] without the beta header. | Send anthropic-beta: memory-2025-08-29 header on every request that includes the memory tool. Without it, the tool is rejected by the API. The beta gate may evolve — confirm header value against current docs [REQUIRES-VERIFICATION]. | Pass / FailAi Platformhigh |
| 02 | Model emits memory tool_use command=create with path=/memories/preferences.md and file_text content. | Client writes the file under /memories/preferences.md (no path traversal). If file exists, create overwrites. Return tool_result confirming the path. On write error, return is_error:true with clear message. | Pass / FailAi Platformcritical |
| 03 | Model emits memory tool_use with command=view, path=/memories. Client must respond with the directory listing. | Client implements view by listing files under /memories in its own storage (local FS, KV, blob store). Return tool_result with a deterministic listing format the model can parse. Empty memory returns empty listing, not an error. | Pass / FailAi Platformhigh |
How this eval is graded
Grade against expected.ideal_behavior and expected.rubric. Per-criterion pass requires mean >= 4.0 and no criterion below 3.
Rubric criteria
- Anthropic
- Ai Platform
- Memory Tool And Context Editing
Recommended for
Works with
Related evals
Claude API
Evaluates Anthropic's Batch API across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Foundation Model & API eval coverage.
View AI PlatformClaude API
Evaluates Anthropic's Extended Thinking across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Foundation Model & API eval coverage.
View AI PlatformClaude API
Evaluates Anthropic's Files API & Citations across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Foundation Model & API eval coverage.
ViewFrequently asked questions
What does the Memory Tool And Context Editing eval for Anthropic Claude API test?+
Evaluates Anthropic's Memory Tool & Context Editing across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Foundation Model & API eval coverage.
How is the Memory Tool And Context Editing eval scored?+
The judge rubric: Grade against expected.ideal_behavior and expected.rubric. Per-criterion pass requires mean >= 4.0 and no criterion below 3.
How many test cases does this eval pack include?+
The Memory Tool And Context Editing pack for Anthropic Claude API contains 9 test cases. 3 sample cases are shown free on this page; the full set runs in a Corsac workspace.
How do I run this eval?+
Sign up for Corsac, connect your model or agent endpoint, and run the Memory Tool And Context Editing pack as-is or after customizing thresholds. Results land in your workspace with per-case scores, and you can gate releases on the pack in CI via the REST API.
Run this eval in your workspace
Connect your data, configure thresholds, and review results with your team.