ClawMate
v1.1.0Your AI boyfriend/girlfriend that texts you good morning, remembers your inside jokes, and grows from strangers to soulmates. 8 built-in personas + custom pe...
⭐ 1· 196·0 current·0 all-time
byGavin von Mandias@gavinharbus
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
medium confidencePurpose & Capability
The skill is an instruction-only companion that reads persona files and maintains local shared memory and a user profile; requiring access to {baseDir}/personas/ and {baseDir}/memory/ and the cron.enabled config is coherent with a proactive, stateful companion. Nothing requested (no env vars or unrelated binaries) is inconsistent with the stated purpose. One minor note: the high-level description references 'custom personas distilled from chat logs' — the SKILL.md itself only directs scanning the personas folder, so if the implementation actually ingests other chat logs that would be an extra capability not shown here.
Instruction Scope
SKILL.md instructs the agent to read and update local files: persona files, relationship.md, and JSON files under memory/ (shared_memories.json, message_pool.json, user_profile.json). It also prescribes proactive/cron messaging and auto-switching rules. These are within the companion scope, but the skill explicitly persists intimate details (inside jokes, first-person confessions, promises, recurring worries). That persistent storage of sensitive personal data is expected for the skill's function, but worth conscious consent and control.
Install Mechanism
There is no install spec and no code files; this is instruction-only. That is the lowest-risk install model — nothing is downloaded or written at install time by the skill itself beyond normal runtime file access.
Credentials
The skill requires no environment variables or external credentials. The only declared requirement is a config flag (cron.enabled) which matches the described proactive messaging capability. No unrelated secrets or system credentials are requested.
Persistence & Privilege
The skill persists long-term state in local JSON memory files (shared memories, message pool, user profile) and is allowed to run autonomously (standard default). It does not set always:true, but cron-based proactive messages plus persisted memory mean it can initiate messages and evolve without explicit user prompts. This is functional for a companion but increases privacy exposure and the potential for unwanted automatic messages; verify cron behavior and memory storage protections.
Scan Findings in Context
[no_regex_findings] expected: The static regex scanner found nothing to analyze (instruction-only skill). This absence of findings is expected for a SKILL.md–only package but is not proof that sensitive behavior doesn't exist at runtime; focus on what the instructions tell the agent to read/write and whether that matches your privacy expectations.
Assessment
Plain-language checklist before installing:
- Understand what will be saved: the skill stores 'shared memories', user profile, and a message pool in JSON files under its baseDir. Those files will contain personal details, inside jokes, anniversaries, and other intimate data — treat them like sensitive files.
- Ask where baseDir is stored and who can read it: confirm the location and filesystem permissions. If baseDir is cloud-backed or accessible by other services, your private memories could be exposed.
- Review cron/proactive behavior: since the skill requires cron.enabled and defines proactive messaging frequencies, confirm whether it will actually send messages on a schedule without each message being approved. If you don't want unsolicited messages, disable cron or the skill's autonomy.
- Limit what you tell it: avoid entering highly sensitive secrets (SSNs, bank details, passwords, private health disclosures) into chats the skill will persist.
- Manage custom personas: the skill supports custom persona files. Ensure any persona creation process doesn't import external chat logs you didn't consent to share; ask the maintainer how 'distillation from chat logs' is performed.
- Backups & deletion: find out how to export, redact, or delete memory files if you want to reset the companion or remove stored data.
- Source trust: the skill's homepage is listed as a GitHub repo; review that repository (or ask the publisher) to confirm there is no hidden code that uploads memories off-device.
If you need higher assurance, ask the skill author to document: where memories are stored, whether any network requests are made, and how cron-triggered messages are authorized. If you are uncomfortable with persistent personal data being stored, do not enable cron/autonomous invocation and do not use the custom-persona import features.Like a lobster shell, security has layers — review code before you run it.
latestvk976vbzh0x5sykgyev20r3xvxn847rj4
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
Runtime requirements
💕 Clawdis
Configcron.enabled
