Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
lossless-claw
v1.0.1无损回忆技能。对对话或会话记录做本地蒸馏，提取身份信息、偏好、任务和长期知识，剔除噪声并保留可追溯日志。
⭐ 3· 3k·12 current·13 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The name/description (local, lossless distillation of conversation logs) aligns with the SKILL.md workflow (read session logs, extract identity/preferences/tasks, promote to long-term). However the recommended commands reference a Python script (scripts/personal_ai_memory.py) that is not included in the skill bundle, which is an inconsistency: the skill tells the agent to run a local tool that doesn't exist in the package.
Instruction Scope
Instructions explicitly tell the agent to 'read session text or logs' and give example CLI usage that accepts arbitrary filesystem paths. There are no limits or guardrails on which paths to read, no explicit list of allowed input types, and no specification of what the --apply flag does—this open-ended file access and potential write/persistence is broad and could lead to processing sensitive files outside the skill's intended scope.
Install Mechanism
This is an instruction-only skill with no install spec and no code files—lowest install risk. The absence of an install step means nothing will be written or executed by default from the package itself, but the provided example commands assume external tooling.
Credentials
The skill declares no required environment variables, credentials, or config paths, which is proportionate to an instruction-only memory-distillation helper. Note: despite no credentials requested, the instructions encourage reading local files which can contain sensitive data—this is a scope but not a credential mismatch.
Persistence & Privilege
The skill is not set to always:true and is user-invocable (normal). However the recommended --apply usage implies making persistent changes (promoting content to a long-term layer) without declaring where that long-term storage lives or what permissions/writes will be performed—this lack of clarity raises concern about untracked persistence.
What to consider before installing
Before installing or enabling this skill, consider: (1) The SKILL.md expects a local script (scripts/personal_ai_memory.py) that is not provided—do not run unreviewed scripts; obtain and inspect any referenced code. (2) The instructions allow reading arbitrary files—limit the skill to explicit, audited conversation files and avoid granting it broad filesystem access. (3) Clarify where 'long-term' data is stored and who can access it; avoid running --apply until storage location and format are known. (4) Test with non-sensitive data first and run in an isolated environment if possible. (5) If you plan to use a third-party implementation of the referenced script, review its source for exfiltration or unwanted writes before giving it access to your logs.Like a lobster shell, security has layers — review code before you run it.
latestvk978jymjgfsghd18jrktg0t74183ydd2
License
MIT-0
Free to use, modify, and redistribute. No attribution required.