Public scan — anyone with this URL can view this analysis. Sign up to track your own repos privately, run scheduled re-scans, and get AI fix prompts via your dashboard.

Panniantong/Agent-Reach

https://github.com/Panniantong/Agent-Reach · scanned 2026-07-23 19:49 UTC (4 days, 20 hours ago)

12 raw signals (0 security + 12 graph)

File as issue Image

UNIFIED Repobility · multi-layer engine · AI coders

Complete repo analysis

Last scanned 4 days, 20 hours ago · v7 · 12 actionable findings from 1 signal source. Security checks, system graph analysis, and verified AI-agent feedback are merged into one review queue.

JSON

70.0% cov · 12 findings

Severity distribution — click a segment to filter

Active filters: excluding tests × Reset all

Severity: Critical 0 High 1 Medium 7 Low 3 9-Layer: Software Security Quality Integrity Frontend Hardware Data Network Cicd Source: Security checks 0 System graph 12 Crowd 0 Layer: Security 8 Quality 3 Cicd 1

Exclude dismissed / FP Exclude test files

Corpus Intelligence Cross-corpus context (cohort percentile, top patterns, fix plan) is shown only on repositories you own. Sign up and connect your repo to view it.

Scan summary Repository scanned at 72.3/100 with 70.0% coverage. It contains 587 nodes across 0 cross-layer flows, written primarily in mixed languages. Engine surfaced 12 findings — concentrated in security (8), quality (3), cicd (1). Risk profile is high: 0 critical, 1 high, 7 medium. Recommended next step: open the security layer findings first — that's where the highest-impact wins live.

Showing 10 of 12 actionable findings. Click TP / FP to vote on a finding's accuracy — votes adjust the confidence weighting and improve detection across the platform.

high System graph security Skillspector conf 0.85 SkillSpector TP2 (tool-poisoning) in agent_reach/skill/SKILL.md

MCP Tool Poisoning Mixing characters from multiple Unicode scripts in a single identifier is a common technique to create visually ambiguous tool names. Skill: agent-reach Rule: TP2 Category: tool-poisoning Severity: HIGH Confidence: 0.85 Remediation: Restrict identifier fields to a single Uni…

agent_reach/skill/SKILL.md:1 Mcp skillTool poisoningTp2

medium System graph quality Agent instructions conf 1.00 Agent authority lacks a verifier contract: agent_reach/skill/SKILL.md

This agent instruction grants code or shell authority but does not state the verification gate that decides promotion. The recurring safe pattern is: LLM proposes; deterministic tests/build/security checks verify; only verified code promotes.

agent_reach/skill/SKILL.md VerificationSkill file

medium System graph security Semgrep conf 0.55 dynamic urllib use detected — agent_reach/channels/bilibili.py:28

Detected a dynamic value being used with urllib. urllib supports 'file://' schemes, so a dynamic value controlled by a malicious actor may allow them to read arbitrary files. Audit uses of urllib calls to ensure user data cannot control the URLs, or consider using the 'requests' library instead. R…

agent_reach/channels/bilibili.py:28 SecurityPython

medium System graph security Semgrep conf 0.55 dynamic urllib use detected — agent_reach/channels/v2ex.py:16

agent_reach/channels/v2ex.py:16 SecurityPython

medium System graph security Semgrep conf 0.55 dynamic urllib use detected — agent_reach/channels/web.py:33

agent_reach/channels/web.py:33 SecurityPython

medium System graph security Coverage conf 1.00 No auth library detected

The scanner did not find any standard auth library (JWT, OAuth, NextAuth, Auth0, etc.). The repo has auth/admin/session surface indicators, so auth may live in custom code, in a separate service, or be missing.

auth

medium System graph security Skillspector conf 0.50 SkillSpector E1 (data-exfil) in agent_reach/skill/references/video.md

https://api.bilibili.com/ Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended. Skill: agent-reach Rule: E1 Category: data-exfil Severity: MEDIUM Confidence: 0.50 Remediation: Verify the destination URL is trusted and nece…

agent_reach/skill/references/video.md:92 Mcp skillData exfilE1

medium System graph security Skillspector conf 0.60 SkillSpector RA2 (rogue-agent) in agent_reach/skill/SKILL_en.md

create files in the agent workspace.** Use `/tmp/` for temporary output and `~/.agent-reach Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction. Skill: a…

agent_reach/skill/SKILL_en.md:110 Mcp skillRogue agentRa2

low System graph quality Debug conf 1.00 Debug logging residue appears in source files

Found 317 console/debugger/print-style debug statements in non-test source. This is a common fast-generation residue before production cleanup.

CleanupRepo hardeningGenerated repo pattern

low System graph quality Complexity conf 1.00 Very large file: agent_reach/cli.py (1834 lines)

Files with >800 lines often hide complexity hotspots and discourage tests.

For AI agents: Voting guide (TP/FP) MCP manifest Stdio wrapper SARIF Integrate Findings queue Vote TP/FP on findings to calibrate the engine.

For AI agents + API integrations

Email me when this repo regresses

Free. We re-scan periodically; new criticals → your inbox. No signup required for the scan itself.

API access

This page is publicly accessible at: https://repobility.com/scan/c7f81467-3935-445f-9c22-47ec0c2d13e0/

To check status programmatically (no auth required):

curl -s https://repobility.com/api/v1/public/scan/c7f81467-3935-445f-9c22-47ec0c2d13e0/

Important — please don't re-submit the same URL repeatedly. The submission endpoint is idempotent: re-submitting the same git URL returns this same scan_token, not a new one. To re-scan this repo, sign up free and use the dashboard.