agentsealgetagentseal/agentseal
Scans AI agents, MCP tools and RAG pipelines for security issues and runs red-team tests to surface prompt injection,…
No ratings yet
Rate this skill
Stars377
Forks68
Watchers377
Open Issues9
InstallOne-line install
UnconfirmedWhether this repository follows the Agent Skills spec (that is, whether it contains a SKILL.md) has not been verified yet. Until that is settled, clone it and use what you need:
git clone https://github.com/getagentseal/agentseal.git
OverviewOverview
Key Features
Scans AI agents, MCP tools and RAG pipelines for security issues and runs red-team tests to surface prompt injection, tool poisoning and credential theft risks early. Key capabilities include: 1. Pattern signatures: ships a library of known malicious patterns (credential access, data exfiltration, etc.) and scans agent configs and tool definitions to flag high-risk items directly; 2. Deobfuscation: decodes Unicode tag characters and encoded obfuscation to expose hidden payloads and disguised malicious instructions; 3. Semantic analysis: judges instructions in context rather than by keywords alone to detect privilege abuse or manipulation and reduce false negatives; 4. MCP tool poisoning probes: specifically probes for malicious instructions injected via MCP tool descriptions and return values, preventing tool-side hijacking of the agent; 5. RAG poisoning probes: checks knowledge base documents for planted misleading content so retrieval results cannot contaminate agent output; 6. Baseline tracking: records historical agent behavior baselines and alerts on deviations to catch gradual tampering; 7. Registry enrichment and custom rules: enriches findings with external registry intelligence and supports custom rules for team-specific policies.
Use Cases
Target Roles
AI / Algorithm Engineer, Network & System Security Engineer, DevOps / Operations Engineer, Test / QA Engineer.
TopicsTopics
RatingUser ratings
Nobody has rated this skill yet. Yours would be the first on the site — your vote becomes the initial average.
RepositoryRepository
| Skill ID | 2101 |
|---|---|
| Data source | GitHub |
| Repository | getagentseal/agentseal |
| Category | AI & Agent Skills / AI Security & Red Teaming |
| Primary language | Python |
| License | Other |
| Default branch | main |
| GitHub repository ID | 1171508227 |
| Node ID | R_kgDORdPMAw |
| Rank | #1597 (position within the source ranking) |
| Site-wide popularity rank | #1798 (computed live from Stars, across all 5833 entries) |
GrowthStar / Fork growth
One snapshot per day, tracking how Stars and Forks changed
- Now377
- Previous point · 09-24 375
- vs previous point +2
- 2 data points 17 days
- Now68
- Previous point · 09-24 69
- vs previous point -1
- 2 data points 17 days
TimelineTimeline
Repository created2026-03-03 10:01:588 months ago
Data last updated2026-10-26 12:16:5423 days ago
Last commit2026-06-11 10:17:585 months ago
CommitLatest commit
Version [latest] committed by Miles-dev-29. Update Category: Feature Addition & Enhancement. Commit details: Merge pull request #32 from getagentseal/feat/dynamic-agent-scan. This update introduces new capabilities and module extensions.