Security audit

Seo Reporter

Security checks for vulnerabilities and agentic risk

Overview

This SEO skill appears purpose-built and not malicious, but it can make unrestricted web requests from the user's environment and needs review before use.

Review this skill before installing or running it. Use it only on public websites you intend to audit, because it sends requests from your agent's environment to the provided URL plus related robots and sitemap paths, follows redirects, and does not prevent access to internal or private network addresses.

A.I.G

Vulnerability Patterns

Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
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

Findings (2)

T09 · Insecure Skill Coding Practices

Error

Location: audit.sh:44
Finding: Unrestricted URL Fetching Enables Server-Side Request Forgery
Content: View full analysis
/dev/null || echo "000") ``` Additional requests are derived from the untrusted target host: ```bash DOMAIN=$(echo "$TARGET_URL" | sed -E 's|https?://([^/]+).*|\1|') PROTOCOL=$(echo "$TARGET_URL" | sed -E 's|(https?)://.*|\1|') BASE_URL="${PROTOCOL}://${DOMAIN}" ROBOTS_URL="${BASE_URL}/robots.txt" SITEMAP_URL="${BASE_URL}/sitemap.xml" ``` ```bash ROBOTS_STATUS_CODE=$(curl -s -o /dev/null -w "%{http_code}" --max-time 10 "$ROBOTS_URL" 2>/dev/null || echo "000") ROBOTS_CONTENT=$(curl -s --max-time 10 "$ROBOTS_URL" 2>/dev/null || echo "") ``` ```bash SITEMAP_STATUS_CODE=$(curl -s -o /dev/null -w "%{http_code}" --max-time 10 "$SITEMAP_URL" 2>/dev/null || echo "000") SITEMAP_CONTENT=$(curl -s --max-time 15 "$SITEMAP_URL" 2>/dev/null || echo "") if [[ "$SITEMAP_STATUS_CODE" == "404" ]]; then # Try common alternatives SITEMAP_ALTS=( "${BASE_URL}/sitemap-index.xml" "${BASE_URL}/sitemap_index.xml" "${BASE_URL}/sitemap/index.xml" ) FOUND_SITEMAP="" for ALT in "${SITEMAP_ALTS[@]}"; do if [[ $(curl -s -o /dev/null -w "%{http_code}" --max-time 10 "$ALT" 2>/dev/null || echo "000") == "200" ]]; then FOUND_SITEMAP="$ALT" break fi done ``` ### Technical Analysis The script accepts an arbitrary user-provided URL and only verifies that its textua ...[truncated 2445 chars]
Remediation: View remediation

T09 · Insecure Skill Coding Practices

Warning

Location: audit.sh:62
Finding: Unbounded Remote Responses and Unescaped Report Fields Permit Resource Exhaustion and Output Manipulation
Content: View full analysis
/dev/null || echo "000") ``` The robots and sitemap response bodies are loaded into shell variables without size limits: ```bash ROBOTS_STATUS_CODE=$(curl -s -o /dev/null -w "%{http_code}" --max-time 10 "$ROBOTS_URL" 2>/dev/null || echo "000") ROBOTS_CONTENT=$(curl -s --max-time 10 "$ROBOTS_URL" 2>/dev/null || echo "") ``` ```bash SITEMAP_STATUS_CODE=$(curl -s -o /dev/null -w "%{http_code}" --max-time 10 "$SITEMAP_URL" 2>/dev/null || echo "000") SITEMAP_CONTENT=$(curl -s --max-time 15 "$SITEMAP_URL" 2>/dev/null || echo "") ``` Extracted remote values are inserted into the generated Markdown without escaping or control-character filtering: ```bash ## Title Tag **Current:** ${TITLE_RAW:-""} $(for f in "${TITLE_FINDINGS[@]:-}"; do echo "- $f"; done) --- ## Meta Description **Current:** ${META_DESC:-""} $(for f in "${META_DESC_FINDINGS[@]:-}"; do echo "- $f"; done) --- ## H1 & Heading Structure **First H1:** ${H1_FIRST:-""} $(for f in "${H1_FINDINGS[@]:-}"; do echo "- $f"; done) $(for f in "${HEADING_STRUCT_FINDINGS[@]:-}"; do echo "- $f"; done) --- ## Canonical Tag **Canonical:** ${CANONICAL:-""} $(for f in "${CANONICAL_FINDINGS[@]:-}"; do echo "- $f"; done) ``` Open Graph values are rendered in a Markdown table without sanitization: ```bash ## Open Graph Tags | Tag | Value | |-----|-------| | og:title | ${OG_TITLE:-""} | | og:description | ${OG_DESC:-""} | | og:image | ${OG_IMAGE:-""} | | og:url | ${OG_URL:-""} | ``` ### Technical Analy ...[truncated 2383 chars]
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (3)

Undeclared Tool Scope

Medium

Category: MCP Least Privilege
Confidence: 88% confidence
Finding: The skill instructs the agent to execute shell commands (./audit.sh and curl is required) but does not declare any tool scope restrictions such as permissions or allowed-tools. That creates an unnecessary trust gap: an agent or runtime may permit broader shell access than is actually needed, increasing the blast radius if the skill is modified, misused with attacker-controlled input, or if helper scripts perform unsafe operations.
Content: No source excerpt is available for this finding.

External Transmission

Medium

Category: Data Exfiltration
Confidence: 93% confidence
Finding: The curl invocation transmits network requests to attacker-controlled destinations based on untrusted input. Because redirects are enabled with -L and no destination validation is performed, an attacker could use this to scan reachable network space, hit internal-only services, or access sensitive metadata endpoints from the agent's environment.
Content: Scanner excerpt · audit.sh (reported line 61)May include surrounding context.
sh
log "Auditing: ${TARGET_URL}" # ── Fetch ───────────────────────────────────────────────────────────────────── HTTP_CODE=$(curl -s -o "$HTML" -D "$HEADERS" -L -A "seo-reporter/1.0" \ -w "%{http_code}" --max-time 20 "$TARGET_URL" 2>/dev/null || echo "000") if [[ "$HTTP_CODE" == "000" ]]; then

Missing User Warnings

Low

Category: Not specified by scanner
Confidence: 87% confidence
Finding: The script performs outbound requests to a fully user-supplied URL and then additionally probes derived resources on the same host such as /robots.txt and /sitemap.xml. In an agent context, this creates an SSRF-style capability and a privacy risk because the skill can be induced to contact internal services, cloud metadata endpoints, or other sensitive network locations without any restriction or user warning.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.