Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/ofershap/conversation-memory/recallgit clone --depth 1 https://github.com/ofershap/conversation-memoryWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00012 | $0.00342 |
| Opus 5 | $0.00006 | $0.00171 |
| Sonnet 5 | $0.00002 | $0.00068 |
| Haiku 4.5 | $0.00001 | $0.00034 |
Grade A, and why
recall scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
Search your conversation history to find and recall context from past sessions.
Use the conversation-memory skill for history file locations and search strategies.
When a search query is provided
-
Determine which AI tool's history to search:
- If in Claude Code: search
~/.claude/history.jsonlfirst - If in Cursor: search
~/.cursor/projects/*/agent-transcripts/ - If unclear: search both
- If in Claude Code: search
-
Search by keyword using grep/rg — never read entire history files
-
For each matching session, show:
- Date and time
- Project name
- First 80 characters of the conversation topic
- Session ID or transcript path
-
Ask which session to load if multiple matches
-
Read the relevant portion of the selected session
-
Summarize: what was done, what was decided, what's still open
When no search query is provided
Guide the user:
What do you want to recall from a past conversation?
- a) Recent work — show my last 10 sessions across all projects
- b) Specific topic — tell me what to search for (e.g., "auth bug", "migration")
- c) Specific date — tell me roughly when (e.g., "last Tuesday", "two weeks ago")
- d) Specific project — which project's history should I search?
- e) Continue previous task — find where I left off on something
Then proceed with the appropriate search strategy.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 44 lines · 12 tokens per session scan A 73f427f515e0
recall is a command published in the GitHub repository ofershap/conversation-memory (1 stars, last pushed 6mo ago), licensed MIT. It adds 12 tokens to every session and 342 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
sum
Distill insights from this conversation into memories.
search
Search Nowledge Mem for relevant memories.
save
Save current session to Nowledge Mem.
status
Check Nowledge Mem connection and server status.
nowledge-distill-memory
Distill durable insights from this conversation into memories.
nowledge-read-working-memory
Read today's Working Memory briefing.