# AI use cases for Agent Memory (tigerless-labs/agent-memory) Repo: [tigerless-labs/agent-memory](https://github.com/tigerless-labs/agent-memory) | Use cases: 5 | Page: https://dedupedai.com/use-cases/tigerless-labs-agent-memory/ agent-memory is a Python runtime that gives AI agents long-term memory stored as plain Markdown files, with local ranked retrieval and a sleep-time consolidation layer, and Claude Code and Codex share one store. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Share one memory of decisions across Claude Code and Codex: https://dedupedai.com/use-cases/tigerless-labs-agent-memory/share-decisions-between-claude-code-and-codex/ - Recover why a past decision was made and what it said then: https://dedupedai.com/use-cases/tigerless-labs-agent-memory/recover-why-a-past-decision-was-made/ - Correct a stale memory without losing its history: https://dedupedai.com/use-cases/tigerless-labs-agent-memory/correct-a-stale-memory-without-losing-history/ - Rebuild a broken memory index from the Markdown files: https://dedupedai.com/use-cases/tigerless-labs-agent-memory/rebuild-memory-index-from-plain-files/ - Review what the sleep-time pass changed and proposed: https://dedupedai.com/use-cases/tigerless-labs-agent-memory/review-overnight-consolidation-runs/ ## How to share one memory of decisions across Claude Code and Codex with Agent Memory Repo: [tigerless-labs/agent-memory](https://github.com/tigerless-labs/agent-memory) | Difficulty: Intermediate | Time: 60 minutes | Basis: From the README Let Claude Code and Codex read and write the same store so past decisions carry into new sessions. **Who it is for:** A developer who runs both Claude Code and Codex on the same project and keeps re-explaining past choices to each one. **The problem:** Each agent session starts empty, so the same decisions get explained again in every tool. **The outcome:** Both agents answer from one store of decision memories, and each answer points to the file it came from. ### Steps 1. Run the setup step described in the README so the host agents use the shared store at AGENT_MEMORY_STORE. 2. Run `mem recall "why files instead of a database"` to see the ranked L0 list of matching memories. 3. Run `mem context "why files instead of a database"` to get the top hits expanded in one call. 4. Open the same question in Codex and check that it returns the same memory paths. ### Prompt to paste into your agent ```text Use the agent-memory repo at to recall memories about with `mem recall` and `mem context`. Return the file paths and one-line abstracts that matched. Verify that the same paths come back when the query is run from Claude Code and from Codex. ``` ### Caveat The README does not give a full setup command in the excerpt, so the first run depends on following the setup section of the README closely. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/tigerless-labs-agent-memory/share-decisions-between-claude-code-and-codex/ ## How to recover why a past decision was made and what it said then with Agent Memory Repo: [tigerless-labs/agent-memory](https://github.com/tigerless-labs/agent-memory) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Pull the reasoning behind an old choice and check what the memory said on a given date. **Who it is for:** An engineer picking up a project after weeks away who needs the reason behind a design choice and whether it changed. **The problem:** Old decisions are hard to reconstruct, and a later edit can hide what was true at the time. **The outcome:** The engineer gets the current memory, the outline of its file, and the version that was valid on the chosen date. ### Steps 1. Run `mem recall "why files instead of a database"` to find the matching memory path. 2. Run `mem read --level outline` to see its headings before opening the full text. (inferred, check the README) 3. Run `mem recall "why files instead of a database" --as-of ` to check what was valid on that date. (inferred, check the README) 4. Run `mem trace` to see the cited raw messages behind the memory when the summary is not enough. ### Prompt to paste into your agent ```text Use the agent-memory repo at to find the memory about , then show its outline and the version valid on . Expected output: the memory path, its outline, and the as-of answer. Verify that the as-of answer differs from the current file when the memory was superseded. ``` ### Caveat The as-of date syntax is not shown in the README excerpt, so the date argument in the step above is inferred and should be checked against the CLI help. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/tigerless-labs-agent-memory/recover-why-a-past-decision-was-made/ ## How to correct a stale memory without losing its history with Agent Memory Repo: [tigerless-labs/agent-memory](https://github.com/tigerless-labs/agent-memory) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Update or replace a wrong memory while keeping the old version for lookup. **Who it is for:** A developer who finds a memory that no longer matches the codebase and wants a clean fix that keeps an audit trail. **The problem:** Editing a memory in place would erase what was believed before, and deleting it would lose the chain. **The outcome:** The stale memory is replaced by a successor, the old file stays in the store, and recall returns the current version. ### Steps 1. Run `mem recall ""` to get the path of the memory to fix. (inferred, check the README) 2. Run `mem correct ` to revise the memory in place when the fix is small. (inferred, check the README) 3. Run `mem record --supersedes ` to create a successor when the meaning changed. (inferred, check the README) 4. Run `mem recall ""` again to confirm the current version is returned. (inferred, check the README) ### Prompt to paste into your agent ```text Use the agent-memory repo at to correct the memory at so it reflects , keeping the old version for as-of recall. Expected output: the new memory path and the superseded status of the old one. Verify that a normal recall returns only the new version. ``` ### Caveat The README excerpt cuts off the exact argument order for correct, record, and supersede, so those commands need checking against the CLI help before use. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/tigerless-labs-agent-memory/correct-a-stale-memory-without-losing-history/ ## How to rebuild a broken memory index from the Markdown files with Agent Memory Repo: [tigerless-labs/agent-memory](https://github.com/tigerless-labs/agent-memory) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Delete the rebuildable index cache and regenerate it without losing any memory. **Who it is for:** An engineer whose search results look wrong or stale after a crash or a copy of the store. **The problem:** Search results may be out of date or broken, and the engineer fears the index holds data that cannot be recovered. **The outcome:** The index is regenerated from the files, and recall works again with no memory lost. ### Steps 1. Stop any agent session that is writing to the store. 2. Run `rm -rf .index/ && mem rebuild` inside the store directory. 3. Run `mem recall "why files instead of a database"` to confirm search returns results again. 4. Run `ls` and `grep` on the store directory to confirm the Markdown files are still present. ### Prompt to paste into your agent ```text Inside the agent-memory store at , run `rm -rf .index/ && mem rebuild`, then run a recall for . Expected output: a non-empty L0 list. Verify that the count of Markdown memory files is the same before and after the rebuild. ``` ### Caveat The README says the rebuild loses no knowledge, but only files outside .index/ and .state/ hold memory, so runtime state in .state/ should be backed up separately if it matters. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/tigerless-labs-agent-memory/rebuild-memory-index-from-plain-files/ ## How to review what the sleep-time pass changed and proposed with Agent Memory Repo: [tigerless-labs/agent-memory](https://github.com/tigerless-labs/agent-memory) | Difficulty: Advanced | Time: 60 minutes | Basis: From the README Check the dream report and dream status to see what consolidation added, updated, or proposed to delete. **Who it is for:** A developer who lets Agent Memory consolidate in the background and wants to approve any deletions before they happen. **The problem:** Unattended consolidation can change memory, and the developer needs to see changes and confirm deletions. **The outcome:** The developer knows which memories moved, which deletions are waiting for confirmation, and what evidence supports each change. ### Steps 1. Run `mem --json dream-status` to see whether a Dream is running, has failed, or is complete. 2. Open the newest file in the dream-reports directory of the store to read what moved and what was proposed. 3. Follow the evidence pointers in the report to the archived provenance files. 4. Run `mem sleep` only when you want a consolidation pass started manually. ### Prompt to paste into your agent ```text Inside the agent-memory store at , run `mem --json dream-status` and read the latest file in dream-reports/. Return the Dream state, the list of moved memories, and any proposed deletions. Verify that no deletion is applied unless it was confirmed. ``` ### Caveat The README describes automatic Dreaming as triggered only by boundary events when no timer is installed, so a quiet store may not run a pass until the next session starts. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/tigerless-labs-agent-memory/review-overnight-consolidation-runs/