23 matches found
skillscript
Safe, reusable automation authored by agents. skillscript.ai • docs TL;DR — npm install -g skillscript-runtime && skillfile init && skillfile dashboard. See Quickstart. What is Skillscript? Skillscript came from asking: what would a Makefile look like if it built skills instead of binaries? The...
MakerChecker
🛡️ MakerChecker The open-source security layer for AI agents. Deny-by-default enforcement, human approvals, and a cryptographically signed audit trail — so your agent runs only what it's granted and provably can't approve its own work. Website · Live Demo · mc scan · Docs Your agents keep running...
OneAlert
OneAlert AI Security OS Open-source autonomous cyber defense for industrial networks AI agents that detect, investigate, hunt, and respond to threats across your OT/ICS infrastructure — with human approval gates and full audit trails. Demo Setup · Features · Quick Start · Architecture Try It Now...
CVE-2026-83817: codewhale-hq CodeWhale: CodeWhale is a DeepSeek and MiMo coding agent for the terminal
CodeWhale is a DeepSeek and MiMo coding agent for the terminal. Prior to 0.9.0, DEFAULTAUTOMATIONAUTOAPPROVE caused AutomationRecord::taskautoapprove to resolve an omitted setting to true, and the automationcreate schema advertised the same insecure default. enqueueruntask converted that omission...
CVE-2026-123812 Dynatrace MCP Server: create_dynatrace_notebook missing the human-approval gate
Dynatrace MCP Server is an MCP server for Dynatrace Observability. Prior to 1.8.7, the createdynatracenotebook registration in src/index.ts performs a persistent document-write action without calling requestHumanApproval, although the project's other write tools use that gate to obtain operator...
CVE-2026-123812 Dynatrace MCP Server: create_dynatrace_notebook missing the human-approval gate
Dynatrace MCP Server is an MCP server for Dynatrace Observability. Prior to 1.8.7, the createdynatracenotebook registration in src/index.ts performs a persistent document-write action without calling requestHumanApproval, although the project's other write tools use that gate to obtain operator...
CVE-2026-123812 Dynatrace MCP Server: create_dynatrace_notebook missing the human-approval gate
Dynatrace MCP Server is an MCP server for Dynatrace Observability. Prior to 1.8.7, the createdynatracenotebook registration in src/index.ts performs a persistent document-write action without calling requestHumanApproval, although the project's other write tools use that gate to obtain operator...
Treating the Model As an Untrusted Component: Deterministic Policy Enforcement, Human Approval and Taint Tracking against Prompt Injection in Tool-Using Agents, with Validation on Open-Weights Models
Tool-using LLM agents act on instructions embedded in the data they read, so prompt injection is a structural rather than a prompt-wording problem. We compare in-prompt defences hardened prompts, spotlighting, LLM detectors with deterministic enforcement outside the model four reference monitors ...
CVE-2026-54141 OpenHuman: Yuanbao non-web channel can bypass human approval to execute `node_exec`
OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before 0.57.13, a remote direct message received through a configured Yuanbao channel is converted into a generic channel AgentTurnRequest that uses the default tool policy without establishing an...
CVE-2026-54141 OpenHuman: Yuanbao non-web channel can bypass human approval to execute `node_exec`
OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before 0.57.13, a remote direct message received through a configured Yuanbao channel is converted into a generic channel AgentTurnRequest that uses the default tool policy without establishing an...
CVE-2026-54141
The OpenHuman agent harness (desktop application and Rust library) is vulnerable to a code injection and missing authorization flaw in versions prior to 0.57.13 . A remote direct message received via a configured Yuanbao channel is converted into an AgentTurnRequest using the default tool policy ...
CVE-2026-54141 OpenHuman: Yuanbao non-web channel can bypass human approval to execute `node_exec`
OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before 0.57.13, a remote direct message received through a configured Yuanbao channel is converted into a generic channel AgentTurnRequest that uses the default tool policy without establishing an...
skillscript v0.39.2
Safe, reusable automation authored by agents. skillscript.ai • docs TL;DR — npm install -g skillscript-runtime && skillfile init && skillfile dashboard. See Quickstart. What is Skillscript? Skillscript came from asking: what would a Makefile look like if it built skills instead of binaries? The...
skillscript v0.39.1
Safe, reusable automation authored by agents. skillscript.ai • docs TL;DR — npm install -g skillscript-runtime && skillfile init && skillfile dashboard. See Quickstart. What is Skillscript? Skillscript came from asking: what would a Makefile look like if it built skills instead of binaries? The...
@dynatrace-oss/dynatrace-mcp-server's create_dynatrace_notebook missing the human-approval gate
Summary A missing human-approval gate on the createdynatracenotebook tool allows a caller to create persistent tenant-visible documents containing arbitrary content including embedded DQL that other users execute when opening the notebook without operator consent. Details dynatrace-mcp-server...
PT-2026-69225
Summary A missing human-approval gate on the create dynatrace notebook tool allows a caller to create persistent tenant-visible documents containing arbitrary content including embedded DQL that other users execute when opening the notebook without operator consent. Details dynatrace-mcp-server...
skillscript v0.36.1
Safe, reusable automation authored by agents. skillscript.ai • docs TL;DR — npm install -g skillscript-runtime && skillfile init && skillfile dashboard. See Quickstart. What is Skillscript? Skillscript came from asking: what would a Makefile look like if it built skills instead of binaries? The...
PT-2026-55225
Name of the Vulnerable Software and Affected Versions FlowDrop versions 0.0.0 through 1.6.0 Description A missing authorization issue allows forceful browsing. The module, which enables testing and running AI-driven workflows via a chat interface, fails to sufficiently re-evaluate human-in-the-lo...
cruxss-bb-agent
CRUXSS Bug Bounty Agent A semi-autonomous bug bounty hunting...
AEGIS: No Tool Call Left Unchecked -- a Pre-Execution Firewall and Audit Layer for AI Agents
AI agents increasingly act through external tools: they query databases, execute shell commands, read and write files, and send network requests. Yet in most current agent stacks, model-generated tool calls are handed to the execution layer with no framework-agnostic control point in between...