Install payload
Install payload is broadly covered in current results.
100% (9/9)
Browse the directory
/
Showing 9 resources for "prompt-injection"
Saved
/ search · [ ] view density · esc clear input
Active
3 trusted · 6 review in this set — compare to see which signals differ.
Trust snapshot
9 results in this view
Claimed
0%(0/9)
3 trust signals differ in this sample: Package trust, Source provenance, Submitter
Signals differ on Package trust, Source provenance, Submitter — add entries to compare before you install.
Rollout signal scan
2 rollout risk signals in current results
Biggest gaps: metadata review, package integrity. 0 entries have 2+ required gaps.
good
Most at-risk entries in this view
- 67%
Prompt-Injection Content Scanner - Claude Code Hook
No required rollout gaps
- 67%
Prompt Injection Defense For Tool Connected Agents
No required rollout gaps
- 67%
Snyk Agent Scan
No required rollout gaps
- 67%
Scrapling MCP Server
No required rollout gaps
- 67%
GitLab CI Claude Automation Agent
No required rollout gaps
Adoption queue
Browse adoption queue · balanced
0/9 visible results are ready for staged adoption under this preset.
ready 0caution 9hold 0
MCP Server Authoring Security Capability Pack Skill
1 blockers: Metadata review
caution
70/100
Request metadata review from maintainers or internal owners.
skills/mcp-server-authoring-security-capability-pack · trust trusted · confidence 83%
MCP Server Security Hardening Skill
1 blockers: Metadata review
caution
70/100
Request metadata review from maintainers or internal owners.
skills/mcp-server-security-hardening · trust trusted · confidence 83%
Prompt Injection Defense Guardrails Skill
1 blockers: Metadata review
caution
70/100
Request metadata review from maintainers or internal owners.
skills/prompt-injection-defense-guardrails · trust trusted · confidence 83%
GitLab CI Claude Automation Agent
1 blockers: Metadata review
caution
50/100
Request metadata review from maintainers or internal owners.
Collect package checksum or signed artifact information.
agents/gitlab-ci-claude-automation-agent · trust review · confidence 67%
Prompt Injection Defense For Tool Connected Agents
1 blockers: Metadata review
caution
50/100
Request metadata review from maintainers or internal owners.
Collect package checksum or signed artifact information.
guides/prompt-injection-defense-for-tool-connected-agents · trust review · confidence 67%
Prompt-Injection Content Scanner - Claude Code Hook
1 blockers: Metadata review
caution
50/100
Request metadata review from maintainers or internal owners.
Collect package checksum or signed artifact information.
hooks/prompt-injection-content-scanner-hook · trust review · confidence 67%
Scrapling MCP Server
1 blockers: Metadata review
caution
50/100
Request metadata review from maintainers or internal owners.
Collect package checksum or signed artifact information.
mcp/scrapling-mcp-server · trust review · confidence 67%
Secure Deployment for Claude Agent SDK Applications
1 blockers: Metadata review
caution
50/100
Request metadata review from maintainers or internal owners.
Collect package checksum or signed artifact information.
guides/secure-deployment-for-claude-agent-sdk-applications · trust review · confidence 67%
Decision confidence
Decision confidence scan · balanced
3/9 results are high-confidence for the selected preset.
high 3medium 6low 0
MCP Server Authoring Security Capability Pack Skill
Confident candidate for staged adoption.
high
74/100
Missing: Metadata review
skills/mcp-server-authoring-security-capability-pack · trust trusted
MCP Server Security Hardening Skill
Confident candidate for staged adoption.
high
74/100
Missing: Metadata review
skills/mcp-server-security-hardening · trust trusted
Prompt Injection Defense Guardrails Skill
Confident candidate for staged adoption.
high
74/100
Missing: Metadata review
skills/prompt-injection-defense-guardrails · trust trusted
GitLab CI Claude Automation Agent
Address Metadata review, Package integrity before broader rollout.
medium
54/100
Missing: Metadata reviewMissing: Package integrity
agents/gitlab-ci-claude-automation-agent · trust review
Prompt Injection Defense For Tool Connected Agents
Address Metadata review, Package integrity before broader rollout.
medium
54/100
Missing: Metadata reviewMissing: Package integrity
guides/prompt-injection-defense-for-tool-connected-agents · trust review
Prompt-Injection Content Scanner - Claude Code Hook
Address Metadata review, Package integrity before broader rollout.
medium
54/100
Missing: Metadata reviewMissing: Package integrity
hooks/prompt-injection-content-scanner-hook · trust review
Scrapling MCP Server
Address Metadata review, Package integrity before broader rollout.
medium
54/100
Missing: Metadata reviewMissing: Package integrity
mcp/scrapling-mcp-server · trust review
Secure Deployment for Claude Agent SDK Applications
Address Metadata review, Package integrity before broader rollout.
medium
54/100
Missing: Metadata reviewMissing: Package integrity
guides/secure-deployment-for-claude-agent-sdk-applications · trust review
Freshness distribution
Mostly fresh with a few aging entries
Median age 53 days; 6 fresh, 3 aging or stale of 9 scanned.
Theme distribution
Results center on security
89% of this view shares the top theme. Leading themes: security, mcp, prompt-injection.
- security89% (8)
- mcp44% (4)
- prompt-injection44% (4)
- claude-code22% (2)
- guardrails22% (2)
- tools22% (2)
- agents11% (1)
- ai-infrastructure11% (1)
32 distinct themes across 9 scanned
PreToolUse hook that scans proposed writes to prompt, agent, rule, markdown, and context files for common prompt-injection phrases before the content is saved into an AI-readable surface.
Build layered defenses against prompt injection, data exfiltration, and unsafe tool execution in AI agent systems.
Defend tool-connected agents against prompt injection using documented Claude Code security practices: MCP trust verification, approval gates, least-privilege tools, untrusted content handling, and human review before side-effect tool calls.
Security scanner from Snyk for discovering local AI agent components, including MCP servers and Agent Skills, and checking them for prompt injection, tool poisoning, tool shadowing, toxic flows, malware payloads, credential handling, and hardcoded secrets.
Secure MCP servers with strict tool boundaries, auth controls, dependency hygiene, and abuse-resistant runtime policies.
Expert MCP capability skill for secure server authoring, tool schema discipline, authorization boundaries, and prompt-injection risk review.
Scrapling MCP Serverby Karim Shoair · submitted by oktofeesh1
MCP server for Scrapling web scraping workflows, including HTTP scraping, browser-backed dynamic fetching, stealth fetching, screenshots, persistent browser sessions, selectors, proxies, and prompt-injection sanitization.
Source-backed agent that operates Claude Code inside GitLab CI pipelines — triaging pipeline failures, generating MR descriptions, running automated code review on diffs, and reporting findings back to merge requests via the GitLab API through a narrow write proxy and explicit prompt-injection boundaries.
A practical walkthrough of securely deploying Claude Agent SDK applications: the prompt-injection threat model, isolation options (sandbox runtime, containers, gVisor, VMs), least privilege, the proxy credential pattern, and filesystem controls.