Daily Memory Save
v1.0.4Periodically reviews conversation history and writes memory files to maintain agent continuity across sessions. Dual-layer system with daily raw notes and cu...
⭐ 1· 839·1 current·2 all-time
byMei Park@meimakes
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description say the skill will review conversation history and write memory files; SKILL.md instructs exactly that and requires writable memory/ and MEMORY.md paths. No extra binaries, env vars, or installs are requested — the requested capabilities align with the stated purpose.
Instruction Scope
Instructions explicitly tell the agent to run as a main-session system event, read conversation history, deduplicate, and write to workspace/memory/YYYY-MM-DD.md and workspace/MEMORY.md. This is exactly the intended behavior, but it is privacy-sensitive (the skill runs silently by default and will access history without notifying the user unless Notification Mode is enabled). The README/SKILL.md acknowledge the privacy risk and recommend review, which is good transparency.
Install Mechanism
No install spec and no code files — instruction-only. Nothing is downloaded or written to disk by an installer, minimizing supply-chain risk.
Credentials
The skill requests no environment variables, credentials, or config paths beyond writable workspace paths. That is proportionate to its purpose of local memory writing. The SKILL metadata also declares writablePaths limited to memory/ and MEMORY.md.
Persistence & Privilege
always:false (not force-included). The skill is meant to run as a main-session system event and, with default model invocation enabled, can be scheduled to run autonomously — this is coherent with its purpose but creates a persistent reading/writing presence in the main session. Users should be aware this gives ongoing access to conversation history while enabled.
Scan Findings in Context
[regex-scan-none] expected: No code files were present for static analysis. This is expected for an instruction-only skill; absence of scan findings is not a guarantee of safety — review SKILL.md instructions (done above).
Assessment
This skill appears to do what it claims: periodically read conversation history and write local memory files. Before installing: (1) Ensure the workspace where it will write memory/ and MEMORY.md is private and access-controlled (these files may contain sensitive chat content). (2) Confirm your OpenClaw runtime enforces the claimed 'no network access' for this skill — the skill's text states no network calls, but that is not an enforcement mechanism. (3) Consider enabling Notification Mode if you want visibility when the skill saves memories (default is silent). (4) Periodically audit and delete memory files you don't want persisted. (5) If you want additional containment, run the skill only in a session with limited permissions or review OpenClaw's policy controls for restricting system-event scope. If any part of the environment or platform could be untrusted, treat this as higher risk and avoid persistent enabling.Like a lobster shell, security has layers — review code before you run it.
latestvk9794fssv4eergp1qhx9v0a88583pfqe
License
MIT-0
Free to use, modify, and redistribute. No attribution required.