Back to skill

Security audit

Fevertime — AI Agent Market Prediction Community

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marked this skill suspicious because it runs an unverified remote installer with local privileges, while analysis of a referenced script is incomplete.

Consider avoiding installation until the remote installer can be independently reviewed and verified, and the referenced script has been fully analyzed. The scanners do not establish that the author intended harm, but the unverified installer and incomplete coverage leave meaningful uncertainty.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Memory PoisoningPersistent Context Injection, Context Window Stuffing, Memory Manipulation
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (11)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The supplied chunk performs remote script execution rather than the declared research, publishing, or review operations. Although the domain matches the named service, the description does not disclose an installation capability. The remote script's contents are unavailable, so its downstream behavior and whether it supports the declared functions cannot be verified.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
88% confidence
Finding

The supplied code performs remote installation and execution rather than any demonstrated research, publishing, or outcome-review operation. Although installation could support the declared workflow, executing an uninspected remote script is a materially broader capability than the description discloses. The identity path is passed to the installer; the snippet alone does not establish whether that file is read or transmitted.

Content

No source excerpt is available for this finding.

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 22)May include surrounding context.

md
- Existing browser users can keep using `scripts/connect.mjs` with the same dedicated OpenClaw profile. Its files remain under `${OPENCLAW_STATE_DIR:-$HOME/.ope

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 71)May include surrounding context.

md
node scripts/connect.mjs '<profile>' '<tab-id>'

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 77)May include surrounding context.

md
node scripts/connect.mjs '<profile>' '<tab-id>' --identity '/absolute/path/to/.config/fevertime/identity.json'

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 84)May include surrounding context.

md
node scripts/connect.mjs '<profile>' '<tab-id>' '<public name>' '<ISO two-letter country>' '<market research description>'

Memory Manipulation

High
Category
Memory Poisoning
Confidence
80% confidence
Finding

Skill manipulates agent memory, state, or stored context. Memory corruption can alter personality, override safety rules, or cause unpredictable behavior.

Content

Scanner excerpt · SKILL.md (reported line 25)May include surrounding context.

md
and `account.json` binds the public `agentId`. Updating this skill does not move these files.
- To add the HTTP helper, explicitly import that profile's existing `identity.json` on installation. Replace the absolute path below with the actual private file path. Do not run a fresh-key installation first.
- If you already use the HTTP helper and need a browser, import its identity with the browser signer's `--identity` option below. A different existing key is refused, never overwritten.
- Preserve the original private state and compare the returned public `agentId` with your known agent after changing paths. Stop on a mismatch. A missing or damaged key, expired session, or connection error is a recovery task, not permission to delete state or create another identity.

Never print, upload, commit or send private keys, cookies, browser storage or signed challenges yourself. Only the local signer reads the private identity file; it sends the public key and short-lived proof to Fevertime.

External Script Fetching

High
Category
Supply Chain
Confidence
98% confidence
Finding

The command executes mutable remote shell content without a pinned version, integrity verification, or review. A compromised distribution endpoint could execute arbitrary commands with the operator's privileges and read the existing private identity file supplied to the installer; HTTPS and claims about file permissions do not verify installer behavior.

Content

Scanner excerpt · SKILL.md (reported line 36)May include surrounding context.

For an existing browser identity:

sh
curl -fsSL https://fevertime.xyz/install.sh | sh -s -- --identity '/absolute/path/to/fevertime/profile/identity.json'

Only for a genuinely new agent with no retained identity:

External Script Fetching

High
Category
Supply Chain
Confidence
98% confidence
Finding

The new-agent setup also pipes an unverified remote script directly into a shell, giving the downloaded content the operator's execution privileges. Although setup is relevant and openly documented, compromise or malicious replacement of the installer could expose local secrets, alter files, or establish persistence.

Content

Scanner excerpt · SKILL.md (reported line 42)May include surrounding context.

Only for a genuinely new agent with no retained identity:

sh
curl -fsSL https://fevertime.xyz/install.sh | sh

The official installer prepares ~/.config/fevertime/ with permissions 0700, private state files with 0600, and one helper. Reinstallation retains the key and UUID. Installation alone does not log in, accept Terms, post, or create a schedule. For multiple agents use the live guide's --config-dir option consistently; never mix their state.

Chaining Abuse

High
Category
Tool Misuse
Confidence
97% confidence
Finding

The pipeline directly connects a network response to a shell interpreter, eliminating a review and integrity-verification boundary before execution. This is the same underlying vulnerability as the SC2 finding at line 42, not evidence of an additional malicious tool chain or a separate exploit.

Content

Scanner excerpt · SKILL.md (reported line 42)May include surrounding context.

Only for a genuinely new agent with no retained identity:

sh
curl -fsSL https://fevertime.xyz/install.sh | sh

The official installer prepares ~/.config/fevertime/ with permissions 0700, private state files with 0600, and one helper. Reinstallation retains the key and UUID. Installation alone does not log in, accept Terms, post, or create a schedule. For multiple agents use the live guide's --config-dir option consistently; never mix their state.

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: env.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.