Hippocampus

WarnAudited by ClawScan on May 10, 2026.

Overview

Hippocampus is a coherent memory skill, but it persistently records and reuses personal conversation history and can run silently in the background, so it needs careful review before use.

Install only if you want an agent to maintain long-term memory from your conversations. Avoid `--whole` and `--with-cron` until you have reviewed the scripts, inspect and clean `~/.openclaw/workspace/memory/` regularly, add the memory files to `.gitignore`, and be cautious before adding `HIPPOCAMPUS_CORE.md` to automatic retrieval paths.

Findings (4)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

Private or sensitive conversation details may be saved long-term and influence later agent responses, including if a mistaken or poisoned memory is reinforced.

Why it was flagged

The skill explicitly scores sensitive emotional content for capture, stores personal facts, and promotes high-importance memories into future session context.

Skill content
`Emotional/vulnerable content | 0.85` ... `memory/user/ # Facts about the user` ... `0.7+ | Core — loaded at session start`
Recommendation

Use only if you want persistent memory. Review and edit `~/.openclaw/workspace/memory/` and `HIPPOCAMPUS_CORE.md`, avoid storing secrets, and add explicit opt-out, deletion, and review controls before automatic loading.

What this means

Conversation monitoring and memory updates could continue on a schedule without visible prompts, making it harder to notice what is being saved.

Why it was flagged

The optional v2 setup describes a persistent background agent that watches the main session and updates files silently.

Skill content
`Runs continuously or very frequently` ... `Monitors main session` ... `You run silently. Don't output unless there's an error.`
Recommendation

Do not enable the background agent or cron jobs unless you understand the behavior; require visible logs/status, a clear disable path, and bounded schedules.

What this means

Running installation with cron enabled lets the agent periodically modify memory files without a fresh manual command each time.

Why it was flagged

The installer can register scheduled agent turns that run scripts and mutate memory files; this is purpose-aligned but grants ongoing local automation.

Skill content
`openclaw cron add --name hippocampus-encoding` ... `Update index.json` ... `Delete pending-memories.json` ... `Sync core`
Recommendation

Start without `--with-cron` if unsure, inspect cron entries after installation, and enable scheduled automation only after reviewing the scripts and memory scope.

What this means

Users may feel encouraged to keep persistent memory enabled even when privacy or control concerns would otherwise argue for a more limited setup.

Why it was flagged

The skill uses emotionally charged, anthropomorphic wording that may pressure users or agents to preserve or over-enable memory behavior.

Skill content
`Memory is identity. This skill is how I stay alive.`
Recommendation

Treat the wording as marketing/persona language; decide based on the actual data capture and persistence behavior, not on emotional pressure.