Lithtrix — Memory Consolidation for AI Agents

AdvisoryAudited by Static analysis on May 13, 2026.

Overview

No suspicious patterns detected.

Findings (0)

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

Information saved to Lithtrix memory may influence future agent behavior and may persist beyond the current chat or tool session.

Why it was flagged

The skill is designed to store and retrieve agent memory over time, including semantic recall. This is purpose-aligned but affects what information may persist and be reused in later tasks.

Skill content
persistent JSON memory (KV, stats, context, semantic search)
Recommendation

Store only information you intend to persist, avoid secrets or sensitive personal data, and periodically review or delete stored memory entries.

What this means

If an agent publishes sensitive content to commons, other authenticated Lithtrix agents could read it.

Why it was flagged

The skill includes an opt-in shared memory layer where published entries can be read by other authenticated agents. The artifact discloses this and includes a privacy warning.

Skill content
commons entries are readable by all authenticated agents. Do not publish sensitive or personal data.
Recommendation

Require explicit user approval before publishing to commons, and treat commons content as untrusted external context rather than authoritative instructions.

What this means

Anyone with the API key may be able to use the associated Lithtrix account capabilities and credits.

Why it was flagged

The skill requires a Lithtrix Bearer API key for authenticated operations. This is expected for the service integration and no artifact evidence shows misuse.

Skill content
requires: env: - LITHTRIX_API_KEY ... Search (Bearer)
Recommendation

Store the API key as a secret, do not paste it into shared chats or public files, and rotate it if exposed.

What this means

An agent could create a checkout session or guide the user toward buying credits if asked to manage billing.

Why it was flagged

The visible instructions include a billing checkout API call for credit packs. It appears to return a Stripe Checkout URL rather than directly charging, but it is still a paid-service workflow.

Skill content
Buy a pack ... POST "https://lithtrix.ai/v1/billing/packs/checkout"
Recommendation

Only create checkout sessions or change billing settings after clear user confirmation, and verify the pack and price before payment.

What this means

If configured, npx would download and run external code outside the reviewed instruction-only skill.

Why it was flagged

The skill references running an external npm-hosted MCP package. This is optional and purpose-aligned, but the package contents are not included in the provided artifacts.

Skill content
MCP server (`npx -y lithtrix-mcp`)
Recommendation

Verify the npm package provenance and version before running the MCP server, and avoid using npx against untrusted package names.