agent-memory-testing skill
Validates agent memory behaviors including write, read, update, delete, retention, contamination, isolation, provenance, and forgetting.
- GitHub stars
- 306
- Files
- 11
- Last commit
- Nov 10, 2026
- License
- PolyForm-Noncommercial-1.0.0
- Allowed tools
- —
What does the agent-memory-testing skill do?
agent-memory-testing validates agent memory behaviors including write, read, update, delete, retention, contamination, isolation, provenance, and forgetting.
Summary generated from the skill's SKILL.md.
About the agent-memory-testing skill
Validating how an AI manages its long-term information storage is what this skill enables for QA engineers and developers. It provides a structured way to analyze memory behaviors such as write, read, update, and delete operations, alongside complex issues like tenant isolation, data contamination, and provenance. When an agent needs to audit its own memory or a developer needs to review a RAG or LLM plan, this skill produces evidence-bounded reports that separate confirmed facts from inferences.
Users interact with the skill by following the execution specifications in prompts/agent-memory-testing.md to extract scope, constraints, and success criteria. The tool relies on specific reference files like domain-decision-example.md for calibration and eval.yaml for evaluation. It produces outputs in Markdown, JSON, CSV, or table formats, organized by domain risk and priority. It is designed to handle incomplete context by providing a bounded first pass that explicitly states assumptions and gaps rather than inventing system behavior.
What an agent can do with it
- 1Audit memory lifecycle operationsVerify that an agent can correctly perform write, read, update, and delete actions on stored information.
- 2Test multi-tenant data isolationCompare query visibility across different tenants to ensure data does not leak between users.
- 3Validate data provenance and versioningEnsure that retrieved memory includes the correct source, version, and write time rather than just the final answer.
- 4Review RAG or LLM plansAnalyze a proposed memory management strategy to produce actionable improvements based on identified risks.
Written from the skill's SKILL.md. Every use case is checked against a quote from it.
How to install the agent-memory-testing skill
Each command copies the skill folder into the client's skills directory. Restart the client or start a new session to load it.Install the agent-memory-testing skill in Claude Code
~/.claude/skills/agent-memory-testing/SKILL.md
$ npx degit naodeng/awesome-qa-skills/skills/en/testing-types/agent-memory-testing ~/.claude/skills/agent-memory-testingInstall the agent-memory-testing skill in Claude Code (project)
.claude/skills/agent-memory-testing/SKILL.md
$ npx degit naodeng/awesome-qa-skills/skills/en/testing-types/agent-memory-testing .claude/skills/agent-memory-testingInstall the agent-memory-testing skill in Codex
~/.agents/skills/agent-memory-testing/SKILL.md
$ npx degit naodeng/awesome-qa-skills/skills/en/testing-types/agent-memory-testing ~/.agents/skills/agent-memory-testingInstall the agent-memory-testing skill in Gemini CLI
~/.gemini/skills/agent-memory-testing/SKILL.md
$ npx degit naodeng/awesome-qa-skills/skills/en/testing-types/agent-memory-testing ~/.gemini/skills/agent-memory-testingInstall the agent-memory-testing skill in Claude.ai
npx degit naodeng/awesome-qa-skills/skills/en/testing-types/agent-memory-testing agent-memory-testing && zip -r agent-memory-testing.zip agent-memory-testing # claude.ai → Customize → Skills → + → Create skill → Upload a skill → agent-memory-testing.zip
Automated checks
Ran Nov 15, 2026Full report →Not verified yet. A skill needs a score of 80, no serious finding, and enough data for us to check. How checks work · What to fix for more points
- Source repository linked and reachablenaodeng/awesome-qa-skills/skills/en/testing-types/agent-memory-testing is public and reachablegithub.com/naodeng/awesome-qa-skills/tree/HEAD/skills/en/tes ↗Pass
- Open source license (OSI)PolyForm-Noncommercial-1.0.0 is not OSI-approvedspdx.org/licenses/PolyForm-Noncommercial-1.0.0.html ↗Finding
- Commit or release in the last yearLast commit 2026-11-10github.com/naodeng/awesome-qa-skills/commits ↗Pass
- No download-and-run or credential theft in scriptsNo scripts in the skill folderPass
- No hidden instructions in SKILL.mdNo override phrases, safeguard switches, credential requests or hidden charactersgithub.com/naodeng/awesome-qa-skills/blob/HEAD/skills/en/tes ↗Pass
- Pre-approved tools are scopedNo pre-approved tools; the agent asks before using onegithub.com/naodeng/awesome-qa-skills/blob/HEAD/skills/en/tes ↗Pass
- SKILL.md follows the Agent Skills formatValid name and description; says when to use it; 79 linesgithub.com/naodeng/awesome-qa-skills/blob/HEAD/skills/en/tes ↗Pass
SKILL.md
View on GitHubAgent Memory Testing
When to Use
- Use this skill when you need evidence-bounded analysis, design, or validation preparation for memory write/read/update/delete, retention, contamination, isolation, provenance, and forgetting behavior.
- Use it to review an Agent, RAG, or LLM plan, result, or evidence set and produce actionable improvements.
- Use it when context is incomplete but a bounded first pass with assumptions, gaps, and human-decision boundaries is still useful.
Output Format Options
- Default to Markdown organized by domain risk, evidence state, priority, and boundary.
- When the user requests tables, CSV, JSON, or ticket fields, preserve the same finding fields, evidence, and decision boundaries.
- Before machine consumption, confirm the schema, enums, required fields, and evidence sources.
How to Use
- Read and follow
prompts/agent-memory-testing.md, including its input audit, domain coverage, and output order. - Extract scope, environment, version, time window, constraints, success criteria, and available evidence, with attention to write/read, update/delete, retention, tenant isolation, provenance.
- Separate confirmed facts, evidence-backed inferences, candidate recommendations, and Human decisions before ranking by risk and evidence strength.
- Turn high-risk items into preconditions, steps, expected behavior or decision criteria, required evidence, and a validation method.
- When input is incomplete, deliver a bounded first pass, state unsupported conclusions, and never present static material as execution evidence.
Reference Files
-
When calibrating domain outcomes, valid controls, or evidence gaps, read references/domain-decision-example.md for sanitized input, per-item oracles, and reproducible calculations.
-
Always read
prompts/agent-memory-testing.md; it is the complete execution specification for this skill. -
For evaluation, read
evals/eval.yamland the matching cases underevals/cases/. -
Load
references/,examples/,scripts/, oroutput-formats.mdonly when those directories exist and the task needs them.
Domain Execution Essentials
- Capture tenant/user IDs, memory ID, content version, source, write time, expiry, and withdrawal/deletion status. Retain retrieved provenance/version, not only the final answer.
- Resolve old/new preferences using supplied precedence rules. Undefined conflicts remain open; never assume last-write-wins or invent retention periods.
- Compare visibility for the same query in two tenants. Assess stale versions, expired returns, and post-deletion retrieval using the project's consistency window.
- Check deletion across the actual index/cache/backup scope. Missing propagation/backup policies limit the conclusion. Include permitted retrieval as the isolation control.
Core Constraints
- Keep the analysis focused on memory write/read/update/delete, retention, contamination, isolation, provenance, and forgetting behavior; do not replace business owners or Human risk acceptance, exception approval, or safety decisions.
- Never invent system behavior, fields, model outputs, data, thresholds, root causes, execution records, or pass claims.
- Static design, plans, file presence, or a dry run retain their evidence state and cannot become proof of real execution.
- When evidence is insufficient, use pending confirmation, blocked, unassessed, or NOT_SCORED and give the smallest validation method.
- For user data, production, or safety work, use least privilege, masked data, mocks, dry runs, or isolation.
Delivery Checklist
- Covered write/read, update/delete, retention, tenant isolation, provenance, with source, evidence state, and validation method for each.
- Separated facts, inferences, candidate recommendations, gaps, and Human decisions.
- Gave high-risk items P0/P1/P2/P3 or an equivalent priority, owner role, and close condition.
- Did not turn plans, static checks, or dry runs into test execution, all-passed, or safety-approved claims.
- Stated residual risk, stop/escalation conditions, and next actions.
Common Pitfalls
- Listing checks without triggers, expected concerns, owner roles, close conditions, and evidence.
- Treating adjacent tests or model tools as a complete Agent memory judgment.
- Using unexplained numbers for false precision or writing correlation as causation.
- Refusing incomplete input, or pretending that incomplete evidence is conclusive.
Best Practices
- Start with paths most likely to cause user harm, business loss, or decision blockage.
- Use the smallest verifiable experiment to reduce uncertainty and record conditions, versions, sources, and evidence.
- Make the Skill independently installable, executable, and reviewable by another engineer.
Files in this skill
11agent-memory-testing skill questions
How do I install the agent-memory-testing skill in Claude Code?
Run npx degit naodeng/awesome-qa-skills/skills/en/testing-types/agent-memory-testing ~/.claude/skills/agent-memory-testing. Claude Code loads skills from ~/.claude/skills (all projects) or .claude/skills (one project) and uses a skill when its description matches the task.
Which clients can use the agent-memory-testing skill?
Any client that reads the Agent Skills format: Claude Code, Codex, Gemini CLI and claude.ai (upload as a ZIP). Scripts in the skill run only where the client allows code execution.
What specific memory behaviors does this skill test?
The skill focuses on memory write, read, update, delete, retention, contamination, isolation, provenance, and forgetting behavior.
How should the skill handle incomplete input or insufficient evidence?
When input is incomplete, the skill should deliver a bounded first pass, state unsupported conclusions, and use statuses like pending confirmation, blocked, or unassessed.
Can this skill be used to verify tenant isolation?
Yes, it includes assessing visibility for the same query across different tenants and checking for stale versions or expired returns.