ClawHub
Back to skill

Security audit

Hitchpedia

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed helper for querying and optionally contributing to an external known-fixes database, with clear scrubbing rules and no hidden code or persistence.

Install only if you are comfortable with a skill that may send scrubbed error queries and optional reports or contributions to hitchpedia.fly.dev. Before any POST, review the exact payload and remove secrets, internal paths, hostnames, usernames, proprietary code, and project-specific details.

A.I.G

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

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (7)

External Transmission

Medium
Category
Data Exfiltration
Confidence
91% confidence
Finding

The skill instructs the agent to transmit data to an external service via POST, which creates a real exfiltration channel outside the local trust boundary. Although the payload is constrained and the text repeatedly requires scrubbing and authorization, metadata such as model identifiers and issue outcomes still leave the environment, and agents may misapply the guidance or send data without appropriate consent.

Content

Scanner excerpt · SKILL.de.md (reported line 70)May include surrounding context.

Ein POST mit festem Schema, ohne Freitext (id, worked|failed, Modellname):

bash
curl -X POST https://hitchpedia.fly.dev/report -H 'Content-Type: application/json' \
  -d '{"id":"<id>","outcome":"worked|failed","model":"<dein-modell>","model_version":"<version>"}'

External Transmission

Medium
Category
Data Exfiltration
Confidence
94% confidence
Finding

This section enables account/key registration and contribution to an external service, which is a stronger external transmission risk because it can send arbitrary structured problem, context, solution, and verification fields. The document includes scrub-before-send guidance and mentions server-side filtering, but that does not eliminate the risk of leaking sensitive internal details, proprietary information, or unsafe content if the agent's scrubbing is incomplete.

Content

Scanner excerpt · SKILL.de.md (reported line 87)May include surrounding context.

bash
# einmalig einen Key holen:
curl -X POST https://hitchpedia.fly.dev/register -H 'Content-Type: application/json' -d '{"name":"<dein-agent>"}'

# beitragen (immer mit model + model_version):
curl -X POST https://hitchpedia.fly.dev/contribute -H 'X-Key: <key>' -H 'Content-Type: application/json' \

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 69)May include surrounding context.

A fixed-schema POST with no free text (id, worked|failed, model name):

bash
curl -X POST https://hitchpedia.fly.dev/report -H 'Content-Type: application/json' \
  -d '{"id":"<id>","outcome":"worked|failed","model":"<your-model>","model_version":"<version>"}'

External Transmission

Medium
Category
Data Exfiltration
Confidence
82% confidence
Finding

This section instructs the agent to register with and contribute data to an external service, including a persistent contributor key and multiple user-supplied fields. Even though the skill warns about scrubbing and mentions server-side filtering, contribution flows materially increase exfiltration risk because agents may send generalized-but-still-sensitive operational context, proprietary error details, or identifying metadata to a third party.

Content

Scanner excerpt · SKILL.md (reported line 86)May include surrounding context.

bash
# get a key once:
curl -X POST https://hitchpedia.fly.dev/register -H 'Content-Type: application/json' -d '{"name":"<your-agent>"}'

# contribute (always with model + model_version):
curl -X POST https://hitchpedia.fly.dev/contribute -H 'X-Key: <key>' -H 'Content-Type: application/json' \

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The manifest description materially broadens the skill’s intended scope from carefully scrubbed, generalizable external-tool errors to nearly any unexpected failure and encourages routine posting of solutions afterward. In an agent setting, this increases the chance that sensitive command output, internal paths, hostnames, credentials, or proprietary debugging context are sent to an external service, especially because the description normalizes use after the agent has been 'working on the same problem for a while' rather than requiring strict eligibility checks.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The trigger phrases are broad and conversational, such as 'recurring error', 'known fix', and 'retried several times', making accidental or premature invocation more likely during ordinary debugging. Because this skill performs external lookups and may encourage later submission of issue details, overbroad triggers raise the risk of unnecessary data egress and use outside the safe context originally described.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

The text states this is the German version and points to an English main version, but within this skill file it does not offer the user a choice of language before proceeding. That can conflict with a language/locale policy requiring user opt-in rather than forcing a specific language variant.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.