UTC

praisonai

105 published advisories; version 4.6.78 or later fixes all of them.

Registry
PyPI
Advisories
105
Latest fix
4.6.78
Exploited in the wild
None listed

Check your machine

Finds praisonai in your agent and MCP configs, installed skills, global npm packages, the npx and uvx caches, uv and pipx tools, Claude Code and Cursor. It runs offline and sends nothing anywhere.

pip install "shieldpi>=0.10.0"
shieldpi agent-config . --installed
Downstream in the Official MCP Registry

1 registered MCP server installs an affected version today, by the dependency versions a fresh install resolves.

  • io.github.MervinPraison/praisonaipraisonai@2.3.42
See the trace →
Provenance

Source: github.com/mervinpraison/praisonai, named in the package metadata (not verified)

0 of the last 10 releases of praisonai were published with build provenance: a signed record of the repository and commit each was built from, which the registry checks before it accepts the release.

  • 4.7.26
  • 4.7.25
  • 4.7.24
  • 4.7.23
  • 4.7.22
  • 4.7.21
  • 4.7.20
  • 4.7.19
  • 4.7.18
  • 4.7.17

Green: published with provenance. Grey: without.

OpenSSF Scorecard has not scored this repository.

Provenance from deps.dev (CC BY 4.0); Scorecard from OpenSSF (CDLA-Permissive-2.0). As of 2026-11-18.

Advisories
AdvisorySeverityAffectedFixed inPublished
PraisonAI vulnerable to Remote Code Execution via Broken AST SandboxCVE-2026-61438High< 4.6.784.6.782026-10-29
PraisonAI: Call API localhost-only authentication bypass via spoofed Host headerCVE-2026-61435High<= 4.6.774.6.782026-10-08
PraisonAI: Prompt-injection defense blocks only when 3+ detector families fire simultaneously; realistic single-vector injections pass throuCVE-2026-60086Medium<= 4.6.774.6.782026-10-08
PraisonAI: Jobs API is unauthenticated by default and allows attacker-controlled webhook SSRFCVE-2026-60091High<= 4.6.774.6.782026-10-08
PraisonAI: SecurityPolicy command/path/import restrictions are completely unenforced by the default SubprocessSandbox backendCVE-2026-60085High<= 4.6.774.6.782026-10-08
PraisonAI: API deploy code generator embeds unescaped YAML fields into Python sourceCVE-2026-61433High<= 4.6.774.6.782026-10-08
PraisonAI: AICoder Arbitrary File Write and Command Execution via LLM Tool CallsCVE-2026-61445Critical<= 4.6.774.6.782026-10-08
PraisonAI: Shell command allowlist bypass via find -exec built-in actionCVE-2026-61434High<= 4.6.774.6.782026-10-08
PraisonAI: MCP HTTP-stream transport is unauthenticated by default, exposing tool enumeration and an unvalidated tool-call surfaceCVE-2026-61427High<= 4.6.774.6.782026-10-08
PraisonAI: ContextGatherer include resolution permits absolute and traversal reads outside the workspaceCVE-2026-61431Medium<= 4.6.774.6.782026-10-08
PraisonAI: PGVector and Cassandra knowledge stores interpolate vector dimensions into DDLCVE-2026-60090Medium<= 4.6.774.6.782026-10-08
PraisonAI: Project custom command templates can read outside-workspace files into model promptsCVE-2026-60088Medium<= 4.6.774.6.782026-10-08
PraisonAI: AgentMail webhook mode accepts forged unsigned message.received events and invokes agentsCVE-2026-61436High<= 4.6.774.6.782026-10-07
PraisonAI: Prompt Injection Defense Bypassed for HIGH-Level ThreatsCVE-2026-61439High<= 4.6.774.6.782026-10-07
PraisonAI: Code Injection via f-string Interpolation in Deploy API Server GenerationCVE-2026-62176Critical<= 4.6.774.6.782026-10-07
PraisonAI: AgentMail webhook lacks signature verification, allowing unauthenticated message injection and sender spoofingCVE-2026-61428High<= 4.6.774.6.782026-10-07
PraisonAI: [Auth Bypass] PraisonAI async Jobs API (`/api/v1/runs`) has no authentication — unauthenticated job execution, result theft, cancCVE-2026-55539High< 4.6.584.6.582026-08-25
PraisonAI has a Browser Server WebSocket origin validation bypass via unanchored regex (patch bypass of CVE-2026-40289 / GHSA-8x8f-54wf-vv92CVE-2026-55536Critical< 4.6.584.6.582026-08-25
PraisonAI serve agents --api-key is ignored, allowing unauthenticated remote agent executionCVE-2026-55534High>= 4.6.34, < 4.6.584.6.582026-08-25
PraisonAI: [Path Traversal] agent tools escape the configured workspace via symlinksCVE-2026-55540High< 4.6.584.6.582026-08-25
PraisonAI: Authentication fail-open in Recipe server allows unauthenticated access when API key or JWT auth is configured without a secretCVE-2026-55533High< 4.6.584.6.582026-08-25
PraisonAI vulnerable to Server-Side Request Forgery via DNS rebinding bypass in webhook_url validationCVE-2026-55535Medium<= 4.6.484.6.582026-08-25
PraisonAI workflow include bypasses tools.py autoload opt-in and executes included recipe codeCVE-2026-55522High>= 3.9.26, < 4.6.584.6.582026-08-25
PraisonAI: Origin-validation bypass (startswith prefix match) enables unauthenticated cross-site request forgery against the PraisonAI MCP HCVE-2026-55532High< 4.6.584.6.582026-08-25
PraisonAI: `--api-key` flag on `praisonai serve` is not properly enforcedCVE-2026-55541High< 4.6.584.6.582026-08-25
PraisonAI: [Auth Bypass] `praisonai serve agents --api-key` is silently ignored — agent-invocation routes (`POST /agents`, `POST /agents/{agCVE-2026-55538High< 4.6.584.6.582026-08-25
PraisonAI: Webhook SSRF via DNS fail-open in `JobSubmitRequest.validate_webhook_url()` — bypass of CVE-2026-40114CVE-2026-55537High< 4.6.584.6.582026-08-25
PraisonAI has an origin validation bypass in MCP HTTP Stream transport that allows browser-mediated unauthenticated tool execution on local CVE-2026-55529Medium< 4.6.584.6.582026-08-25
PraisonAI MCP HTTP server has unauthenticated unbounded session accumulation (memory exhaustion; session TTL never enforced)CVE-2026-55531Medium< 4.6.584.6.582026-08-25
PraisonAI Dynamic Context history and terminal tools read files outside configured storage via path traversalCVE-2026-56833High>= 3.8.1, <= 4.6.584.6.592026-06-18
PraisonAI: Arbitrary File Read/Write via `multiedit` Tool Without Path ValidationCVE-2026-57145Critical< 4.6.614.6.612026-06-18
PraisonAI: Unauthenticated RCE via Jobs API + Approval BypassCVE-2026-57125Critical<= 4.6.484.6.592026-06-18
PraisonAI recipe serve Typer command bypasses the non-localhost authentication guardCVE-2026-56836High>= 4.5.112, <= 4.6.584.6.592026-06-18
PraisonAI: HTTPApproval dashboard renders tool arguments as raw HTML, allowing approval-page XSS to approve dangerous toolsCVE-2026-56840High>= 4.5.2, <= 4.6.584.6.592026-06-18
PraisonAI SandlockSandbox falls back to unrestricted subprocess execution when Landlock is unavailableCVE-2026-57144High>= 4.5.110, < 4.6.614.6.612026-06-18
PraisonAI recipe workflow policy can be bypassed by declaring and YAML-approving dangerous tools outside TEMPLATE.yamlCVE-2026-57142High>= 4.5.87, < 4.6.614.6.612026-06-18
PraisonAI DiscordApproval accepts unrelated channel messages as dangerous-tool approvalsCVE-2026-56832High>= 4.5.2, <= 4.6.584.6.592026-06-18
PraisonAI: AgentOS remains unauthenticated after incomplete fix version and allows remote agent invocationCVE-2026-57116Critical>= 4.2.1, <= 4.6.574.6.592026-06-18
PraisonAI: PRAISONAI_CALL_AUTH=disabled environment variable unconditionally disables authenticationCVE-2026-57132High< 4.6.614.6.612026-06-18
PraisonAI GitHub template cache path traversal allows outside-cache file write and directory deletionCVE-2026-57113High>= 2.6.0, <= 4.6.574.6.592026-06-18
PraisonAI LinearBot processes unsigned webhooks when LINEAR_WEBHOOK_SECRET is missingCVE-2026-56837High>= 4.6.56, <= 4.6.584.6.592026-06-18
PraisonAI: Jobs API exposes agent-execution endpoints with no authenticationCVE-2026-57131Critical< 4.6.594.6.592026-06-18
PraisonAI Code agent tools fail open without a workspace boundaryCVE-2026-56839High<= 4.6.574.6.592026-06-18
praisonai: recipe serve auth middleware silently disables itself when no secret is setCVE-2026-57127Critical<= 4.6.484.6.592026-06-18
PraisonAI dynamic-context artifact tools read arbitrary host files outside artifact storageCVE-2026-56834High>= 3.8.1, <= 4.6.584.6.592026-06-18
PraisonAI A2U incomplete authentication fix leaves current serve command unauthenticated by defaultCVE-2026-57146High>= 4.5.115, < 4.6.614.6.612026-06-18
PraisonAI: Unauthenticated Local File Inclusion via agent_file path in PraisonAI Jobs APICVE-2026-57119High< 4.6.594.6.592026-06-18
PraisonAI: Missing Authentication for Critical Function and Improper Neutralization of Special Elements used in an OS Command ('OS Command ICVE-2026-57124Critical<= 4.6.484.6.592026-06-18
PraisonAI Slack app_mention bypasses configured user/channel authorizationCVE-2026-56835High>= 3.11.0, <= 4.6.584.6.592026-06-18
PraisonAI: Jobs webhook SSRF protection bypass via DNS rebindingCVE-2026-57114High>= 4.5.126, <= 4.6.584.6.592026-06-18
PraisonAI recipe.run_stream skips dangerous-tool policy enforcementCVE-2026-56838High>= 4.5.87, <= 4.6.584.6.592026-06-18
PraisonAI ToolsMCPServer legacy SSE transport accepts attacker Host/Origin and exposes registered toolsCVE-2026-57112High>= 3.10.0, <= 4.6.584.6.592026-06-18
PraisonAI: Compute-bridged file tools allow shell command injectionCVE-2026-57117High>= 4.6.10, <= 4.6.584.6.592026-06-18
PraisonAI: Webhook signature verification skipped (fail-open) when secret unset, allowing forged inbound webhooks (WhatsApp & Linear bots)CVE-2026-57122High<= 4.6.524.6.592026-06-18
PraisonAI vulnerable to sandbox escape via `print.__self__` builtins module leak in `execute_code` (subprocess mode)CVE-2026-47392Critical<= 4.6.394.6.402026-05-29
PraisonAI spider_tools SSRF protection bypass via alternate loopback host encodingsCVE-2026-47390Medium<= 4.6.394.6.402026-05-29
PraisonAI CLI automatically resolves @url mentions in prompt text and can read loopback URLs into model contextCVE-2026-47395Medium<= 4.6.394.6.402026-05-29
PraisonAI: Arbitrary code execution via unguarded `spec.loader.exec_module` in `agents_generator.py` - sibling of CVE-2026-44334CVE-2026-47398High<= 4.6.394.6.402026-05-29
PraisonAI `deploy --type api` emits a Flask server with authentication disabled by defaultCVE-2026-47393Critical<= 4.6.394.6.402026-05-29
PraisonAI call server exposes unauthenticated agent listing, invocation, and deletion when CALL_SERVER_TOKEN is unsetCVE-2026-47396Critical<= 4.6.394.6.402026-05-29
PraisonAI vulnerable to unauthenticated arbitrary file read via MCP workflow.show, workflow.validate, deploy.validateCVE-2026-47394High<= 4.6.394.6.402026-05-29
PraisonAI has an Arbitrary File Write in Python APICVE-2026-47397High<= 4.6.394.6.402026-05-29
PraisonAI's unauthenticated A2A official example can reach real LLM-driven `eval()` tool executionCVE-2026-47391Critical<= 4.6.394.6.402026-05-29
PraisonAI knowledge-store backends interpolate unvalidated collection names into SQL and CQL queriesCVE-2026-44337Medium>= 2.4.1, <= 4.6.334.6.342026-05-11
PraisonAI ships and generates a legacy API server with authentication disabled by default, allowing unauthenticated workflow executionCVE-2026-44338High>= 2.5.6, <= 4.6.334.6.342026-05-11
PraisonAI MCP `tools/call` path-traversal => RCE via Python `.pth` injectionCVE-2026-44336Critical<= 4.6.334.6.342026-05-11
PraisonAI's symlink-extraction bypass of `_safe_extractall` writes outside `dest_dir`CVE-2026-44340High<= 4.6.364.6.372026-05-11
PraisonAI has unsafe tool resolution in `ToolExecutionMixin.execute_tool`: undeclared `__main__` callables executeCVE-2026-44339High<= 4.6.364.6.372026-05-11
PraisonAI has unauthenticated RCE via `tool_override.py` (CVE-2026-40287 patch bypass)CVE-2026-44334High>= 4.5.139, <= 4.6.314.6.322026-05-06
PraisonAI has an incomplete fix for CVE-2026-34935 - OS Command InjectionCVE-2026-41497Critical<= 4.5.1484.5.1492026-04-17
PraisonAI: SQL Injection via unvalidated `table_prefix` in 9 conversation store backends (incomplete fix for CVE-2026-40315)CVE-2026-41496High<= 4.5.1484.5.1492026-04-17
PraisonAI Vulnerable to Implicit Execution of Arbitrary Code via Automatic `tools.py` LoadingCVE-2026-40156High< 4.5.1284.5.1282026-04-10
PraisonAI has Unrestricted Upload Size in WSGI Recipe Registry Server that Enables Memory Exhaustion DoSCVE-2026-40115Medium< 4.5.1284.5.1282026-04-10
PraisonAI Vulnerable to Code Injection and Protection Mechanism FailureCVE-2026-40158High< 4.5.1284.5.1282026-04-10
PraisonAI: Unauthenticated Allow-List Manipulation Bypasses Agent Tool Approval Safety ControlsCVE-2026-40149High< 4.5.1284.5.1282026-04-10
PraisonAI Vulnerable to Server-Side Request Forgery via Unvalidated webhook_url in Jobs APICVE-2026-40114High< 4.5.1284.5.1282026-04-10
PraisonAI Browser Server allows unauthenticated WebSocket clients to hijack connected extension sessionsCVE-2026-40289Critical<= 4.5.1384.5.1392026-04-10
PraisonAI vulnerable to arbitrary file write via path traversal in `praisonai recipe unpack`CVE-2026-40157Critical>= 2.7.2, < 4.5.1284.5.1282026-04-10
PraisonAI Vulnerable to Stored XSS via Unsanitized Agent Output in HTML Rendering (nh3 Not a Required Dependency)CVE-2026-40112Medium< 4.5.1284.5.1282026-04-10
PraisonAI Vulnerable to Decompression Bomb DoS via Recipe Bundle Extraction Without Size LimitsCVE-2026-40148Medium< 4.5.1284.5.1282026-04-10
PraisonAI Vulnerable to Argument Injection into Cloud Run Environment Variables via Unsanitized Comma in gcloud --set-env-varsCVE-2026-40113High< 4.5.1284.5.1282026-04-10
PraisonAI Vulnerable to RCE via Automatic tools.py ImportCVE-2026-40287High<= 4.5.1384.5.1392026-04-10
PraisonAI Vulnerable to Sensitive Environment Variable Exposure via Untrusted MCP Subprocess ExecutionCVE-2026-40159Medium< 4.5.1284.5.1282026-04-10
PraisonAI: Unauthenticated Information Disclosure of Agent Instructions via /api/agents in AgentOSCVE-2026-40151Medium< 4.5.1284.5.1282026-04-10
PraisonAI Vulnerable Untrusted Remote Template Code ExecutionCVE-2026-40154Critical< 4.5.1284.5.1282026-04-10
PraisonAI: Unauthenticated WebSocket Endpoint Proxies to Paid OpenAI Realtime API Without Rate LimitsCVE-2026-40116High< 4.5.1284.5.1282026-04-10
PraisonAI: Hardcoded `approval_mode="auto"` in Chainlit UI Overrides Administrator Configuration, Enabling Unapproved Shell Command ExecutioGHSA-qwgj-rrpj-75xmHigh< 4.5.1284.5.1282026-04-10
PraisonAI has critical RCE via `type: job` workflow YAMLCVE-2026-40288Critical<= 4.5.1384.5.1392026-04-10
PraisonAI: SQLiteConversationStore didn't validate table_prefix when constructing SQL queriesCVE-2026-40315Medium< 4.5.1334.5.1332026-04-10
PraisonAI Vulnerable to OS Command InjectionCVE-2026-40088Critical< 4.5.1214.5.1212026-04-08
PraisonAI Vulnerable to Remote Code Execution via YAML Deserialization in Agent Definition LoadingCVE-2026-39890Critical<= 4.5.1144.5.1152026-04-08
PraisonAI Has Unauthenticated SSE Event Stream that Exposes All Agent Activity in A2U ServerCVE-2026-39889High<= 4.5.1144.5.1152026-04-08
PraisonAI has Template Injection in Agent Tool DefinitionsCVE-2026-39891High<= 4.5.1144.5.1152026-04-08
PraisonAI Has Arbitrary File Write (Zip Slip) in Templates ExtractionCVE-2026-39307High<= 4.5.1124.5.1132026-04-06
PraisonAI recipe registry pull path traversal writes files outside the chosen output directoryCVE-2026-39306High<= 4.5.1124.5.1132026-04-06
PraisonAI Has Path Traversal in FileToolsCVE-2026-35615Critical<= 1.5.1121.5.1132026-04-06
PraisonAI Vulnerable to Arbitrary File Write / Path Traversal in Action OrchestratorCVE-2026-39305Critical<= 4.5.1124.5.1132026-04-06
PraisonAI recipe registry publish path traversal allows out-of-root file writeCVE-2026-39308High<= 4.5.1124.5.1132026-04-06
PraisonAI Has ReDoS via Unvalidated User-Controlled Regex in MCPToolIndex.search_tools()CVE-2026-34939Medium<= 4.5.894.5.902026-04-01
PraisonAI Has Authentication Bypass via OAuthManager.validate_token()CVE-2026-34953Critical<= 4.5.964.5.972026-04-01
PraisonAI Has Second-Order SQL Injection in `get_all_user_threads`CVE-2026-34934Critical<= 4.5.894.5.902026-04-01
PraisonAI: OS Command Injection in MCPHandler.parse_mcp_command()CVE-2026-34935Critical>= 4.5.15, < 4.5.694.5.692026-04-01
PraisonAI Has Missing Authentication in WebSocket GatewayCVE-2026-34952Critical<= 4.5.964.5.972026-04-01
PraisonAI Has Sandbox Escape via shell=True and Bypassable Blocklist in SubprocessSandboxCVE-2026-34955High<= 4.5.964.5.972026-04-01
PraisonAI: SSRF via Unvalidated api_base in passthrough() FallbackCVE-2026-34936High<= 4.5.894.5.902026-04-01

Sources: GitHub Advisory Database, reviewed and malware advisories (CC BY 4.0, https://github.com/advisories); advisories published by Cursor (https://github.com/cursor/cursor/security/advisories); advisories published by OpenClaw (https://github.com/openclaw/openclaw/security/advisories); advisories published by vLLM (https://github.com/vllm-project/vllm/security/advisories); advisories published by Open WebUI (https://github.com/open-webui/open-webui/security/advisories); advisories published by n8n (https://github.com/n8n-io/n8n/security/advisories); advisories published by Flowise (https://github.com/FlowiseAI/Flowise/security/advisories); advisories published by Langflow (https://github.com/langflow-ai/langflow/security/advisories); advisories published by the MCP Python SDK project (https://github.com/modelcontextprotocol/python-sdk/security/advisories); advisories published by MLflow (https://github.com/mlflow/mlflow/security/advisories); CISA Known Exploited Vulnerabilities catalog (CC0). Data as of 2026-11-22. ShieldMap describes how each problem works and how to fix it; it does not publish exploit code. Report a correction to [email protected].