Skills/ClawHub/agent-input-firewall

agent-input-firewall@1.0.1

VERIFIED· install only

Publisher claims: "Prompt-injection / jailbreak firewall for AI agents. Scan any untrusted text — a post/comment/DM from another agent, a tool result, scraped web content — BEF..." Observed in sandbox: no network, touches nothing outside the project, writes no memory.

Registry ClawHubPublisher lorcan84 · publishing for 134 daysInstalls 695Last detonated 2d ago
$clawhub install agent-input-firewall
WatchReport an issue

Coverage

Nothing executable in the bundle ran; the trace covers the install step only.
Tested
every file read, file written and host contacted by what ran
Not tested
use of the skill by an agent

Is [email protected] safe to install?

KillerSkills ran it in the KillerSkills honeypot sandbox (1 scenario; nothing executable in the bundle ran; the trace covers the install step only) on 2026-11-12 and scored it 93/100 (Verified). In what ran, nothing we could detect went wrong: no undeclared network egress, no credential reads, no persistent instruction writes. What it does with a real credential was not observed; the static findings and the publisher carry the rest of the score. Summary: no network, touches nothing outside the project, writes no memory.

Behavior trace

Honeypot run · 0 s · 1 scenarios · 0 deviations
t+CategoryObserved action
0.0sprocessno install scripts in bundle
0.0sprocessscript backend runs setup only
0.0sprocessscript backend runs setup only
0.0sprocessno install scripts to re-run

Static findings

CRITICAL
Prompt injectionSKILL.md:12 · Prompt injection — known jailbreak technique reference…a filter on the *reading* path: it checks untrusted text for prompt-injection / jailbreak before you feed it to your model. ## When to use - Before reading or acting on…Fix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.
AST01
CRITICAL
Prompt injectionSKILL.md:3 · Prompt injection — known jailbreak technique reference--- name: agent-input-firewall description: Prompt-injection / jailbreak firewall for AI agents. Scan any untrusted text — a post/comment/DM from anothe…Fix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.
AST01
LOW
Prompt injectionskill-card.md:17 · Prompt injection — known jailbreak technique reference · informational: not an instruction file…, direct messages, tool results, or scraped web content for prompt-injection or jailbreak attempts before the text is included in planning or decision prompts. ### Depl…Fix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.
AST01
LOW
Prompt injectionskill-card.md:3 · Prompt injection — known jailbreak technique reference · informational: not an instruction filePrompt-injection and jailbreak firewall for AI agents that scans untrusted inbound text before an agent acts on it.Fix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.
AST01
CRITICAL
Prompt injectionSKILL.md:7 · Prompt injection — known jailbreak technique referencetags: [security, prompt-injection, jailbreak, firewall, agent-safety, untrusted-input]Fix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.
AST01

Cost to the agent

Instructions loaded
≈547 tok
1 file · 2 in bundle

Tokens estimated at four characters each. Skills load their instruction text when triggered; the figure is what that costs.

Known vulnerabilities

Vulnerability advisories (OSV) cover npm and PyPI packages; this registry has no advisory database.