PraisonAI has 113 CVEs on record. Disclosures have slowed: 22 in the last 90 days after 70 in the 90 before. The busiest recent month was June 2026 with 38. The median CVSS is 8.0 (high), with 18 rated critical. None have a confirmed exploitation report. The dominant weakness classes are CWE-306 (20) and CWE-22 (14).
CVEs per month
Last 12 months, by publish date
- Exploited share
- 0% vs 1% corpus
- Median CVSS
- 8.0
- Publish → KEV
- —
- Last 90 days
- 22 prev 70
Weakness classes
Products
- praisonai 113
Worst active — by depth score
CVE-2026-56075High· 8.8PraisonAI before 4.5.128 contains an arbitrary shell command execution vulnerability where the UI modules hardcode approval_mode to auto, overriding administrator configuration from PRAISON_APPROVAL_MODE environment variable61CVE-2026-44338High· 7.3PraisonAI ships and generates a legacy API server with authentication disabled by default, allowing unauthenticated workflow execution59CVE-2026-56076High· 8.1PraisonAI before 1.5.128 contains a cross-origin agent execution vulnerability in the AGUI endpoint that allows remote attackers to trigger arbitrary agent execution57GHSA-mhgx-w3w5-2rvcCritical· 10.0Duplicate Advisory: PraisonAI: CodeAgent Executes LLM-Generated Code Without Sandboxing and Leaks All Environment Secrets55CVE-2026-61445Critical· 9.9PraisonAI before 4.6.78 contains arbitrary file write and command execution vulnerabilities in the AICoder component due to missing path validation and command sanitization in LLM tool calls55
PraisonAI vulnerabilities
CVEs affecting PraisonAI, newest first. Open any entry for full detail, references, and exploit status.
113 CVEsRSS
GHSA-wgvq-3jxh-4qg7Medium· 5.5Duplicate Advisory: PraisonAI: Project custom command templates can read outside-workspace files into model prompts
Duplicate Advisory: PraisonAI: Project custom command templates can read outside-workspace files into model prompts
GHSA-q65p-7p84-495cCritical· 9.8Duplicate Advisory: PraisonAI: PGVector and Cassandra knowledge stores interpolate vector dimensions into DDL
Duplicate Advisory: PraisonAI: PGVector and Cassandra knowledge stores interpolate vector dimensions into DDL
GHSA-mhgx-w3w5-2rvcCritical· 10.0Duplicate Advisory: PraisonAI: CodeAgent Executes LLM-Generated Code Without Sandboxing and Leaks All Environment Secrets
Duplicate Advisory: PraisonAI: CodeAgent Executes LLM-Generated Code Without Sandboxing and Leaks All Environment Secrets
GHSA-5r2q-3wg3-57m2High· 7.5Duplicate Advisory: PraisonAI: Prompt Injection Defense Bypassed for HIGH-Level Threats
Duplicate Advisory: PraisonAI: Prompt Injection Defense Bypassed for HIGH-Level Threats
GHSA-4mp6-8448-9vgvMedium· 7.3Duplicate Advisory: PraisonAI: AgentMail webhook lacks signature verification, allowing unauthenticated message injection and sender spoofing
Duplicate Advisory: PraisonAI: AgentMail webhook lacks signature verification, allowing unauthenticated message injection and sender spoofing
CVE-2026-61445Critical· 9.9PraisonAI before 4.6.78 contains arbitrary file write and command execution vulnerabilities in the AICoder component due to missing path validation and command sanitization in LLM tool calls
PraisonAI before 4.6.78 contains arbitrary file write and command execution vulnerabilities in the AICoder component due to missing path validation and command sanitization in LLM tool calls. Attackers can inject malicious prompts throug…
CVE-2026-61439High· 7.5PraisonAI versions before 4.6.78 contain a prompt injection defense misconfiguration where the block threshold defaults to CRITICAL severity, allowing HIGH-level threats to pass through unblocked
PraisonAI versions before 4.6.78 contain a prompt injection defense misconfiguration where the block threshold defaults to CRITICAL severity, allowing HIGH-level threats to pass through unblocked. Attackers can submit single-vector promp…
CVE-2026-61428High· 7.3PraisonAI AgentMail versions before 4.6.78 lack signature verification in webhook mode, allowing unauthenticated attackers to inject messages with spoofed sender addresses
PraisonAI AgentMail versions before 4.6.78 lack signature verification in webhook mode, allowing unauthenticated attackers to inject messages with spoofed sender addresses. Attackers can POST crafted message.received events to the webhoo…
CVE-2026-61426High· 8.6PraisonAI before 1.7.3 contains an insecure default configuration that binds to all interfaces with no API key requirement and wildcard CORS
PraisonAI before 1.7.3 contains an insecure default configuration that binds to all interfaces with no API key requirement and wildcard CORS. Unauthenticated attackers can call GET /api/agents to read agent instructions and system prompt…
CVE-2026-60090Critical· 9.8PraisonAI before 4.6.78 fails to validate the caller-controlled dimension argument in the PGVector and Cassandra knowledge-store create_collection() backends
PraisonAI before 4.6.78 fails to validate the caller-controlled dimension argument in the PGVector and Cassandra knowledge-store create_collection() backends. Although schema, keyspace, and collection-name identifiers are validated, the …
CVE-2026-60088Medium· 5.5PraisonAI before 4.6.78 fails to validate file path references in custom command templates, allowing attackers to read files outside the workspace
PraisonAI before 4.6.78 fails to validate file path references in custom command templates, allowing attackers to read files outside the workspace. Attackers can include path traversal sequences like @../outside_secret.txt or absolute pa…
GHSA-3wrm-pm5v-8vq8Medium· 7.2Duplicate Advisory: PraisonAI: Jobs API is unauthenticated by default and allows attacker-controlled webhook SSRF
Duplicate Advisory: PraisonAI: Jobs API is unauthenticated by default and allows attacker-controlled webhook SSRF
GHSA-wrw8-384c-cg76High· 8.8Duplicate Advisory: PraisonAI: Shell command allowlist bypass via find -exec built-in action
Duplicate Advisory: PraisonAI: Shell command allowlist bypass via find -exec built-in action
GHSA-5c3v-h6hw-4gx7Medium· 5.3Duplicate Advisory: PraisonAI: Prompt-injection defense blocks only when 3+ detector families fire simultaneously; realistic single-vector injections pass through unblocked
Duplicate Advisory: PraisonAI: Prompt-injection defense blocks only when 3+ detector families fire simultaneously; realistic single-vector injections pass through unblocked
GHSA-w37c-cq55-frjpMedium· 5.5Duplicate Advisory: PraisonAI: ContextGatherer include resolution permits absolute and traversal reads outside the workspace
Duplicate Advisory: PraisonAI: ContextGatherer include resolution permits absolute and traversal reads outside the workspace
CVE-2026-61444Critical· 9.1PraisonAI versions before 4.6.78 contain a code injection vulnerability in deploy/api.py where the agents_file parameter is directly interpolated into an f-string without sanitization
PraisonAI versions before 4.6.78 contain a code injection vulnerability in deploy/api.py where the agents_file parameter is directly interpolated into an f-string without sanitization. Attackers can inject arbitrary Python code that exec…
GHSA-fwh2-95jw-g4j6High· 8.8Duplicate Advisory: PraisonAI has Memory State Leakage and Path Traversal in MultiAgent Context Handling
Duplicate Advisory: PraisonAI has Memory State Leakage and Path Traversal in MultiAgent Context Handling
GHSA-x44p-gg67-52fcMedium· 5.5Duplicate Advisory: PraisonAI: Coarse-Grained Tool Approval Cache Bypasses Per-Invocation Consent for Shell Commands
Duplicate Advisory: PraisonAI: Coarse-Grained Tool Approval Cache Bypasses Per-Invocation Consent for Shell Commands
CVE-2026-56076High· 8.1PoCPraisonAI before 1.5.128 contains a cross-origin agent execution vulnerability in the AGUI endpoint that allows remote attackers to trigger arbitrary agent execution
PraisonAI before 1.5.128 contains a cross-origin agent execution vulnerability in the AGUI endpoint that allows remote attackers to trigger arbitrary agent execution. The POST /agui endpoint lacks authentication and hardcodes Access-Cont…
CVE-2026-56074Medium· 5.5PoCPraisonAI before 1.5.128 caches tool approval decisions by tool name only, not by invocation arguments, allowing subsequent execute_command calls to bypass approval prompts
PraisonAI before 1.5.128 caches tool approval decisions by tool name only, not by invocation arguments, allowing subsequent execute_command calls to bypass approval prompts. Attackers can exploit this by obtaining initial approval for a …
CVE-2026-56075High· 8.8PoCPraisonAI before 4.5.128 contains an arbitrary shell command execution vulnerability where the UI modules hardcode approval_mode to auto, overriding administrator configuration from PRAISON_APPROVAL_MODE environment variable
PraisonAI before 4.5.128 contains an arbitrary shell command execution vulnerability where the UI modules hardcode approval_mode to auto, overriding administrator configuration from PRAISON_APPROVAL_MODE environment variable. Authenticat…
GHSA-8579-rgg5-ph2mHigh· 8.8PraisonAI DiscordApproval accepts unrelated channel messages as dangerous-tool approvals
PraisonAI DiscordApproval accepts unrelated channel messages as dangerous-tool approvals
GHSA-22cj-m4wf-fv2cHigh· 7.5PraisonAI Dynamic Context history and terminal tools read files outside configured storage via path traversal
PraisonAI Dynamic Context history and terminal tools read files outside configured storage via path traversal
GHSA-5qw8-f2g9-ff29High· 8.2PraisonAI recipe serve Typer command bypasses the non-localhost authentication guard
PraisonAI recipe serve Typer command bypasses the non-localhost authentication guard
GHSA-qvpf-j64c-jmhrHigh· 8.3PraisonAI Slack app_mention bypasses configured user/channel authorization
PraisonAI Slack app_mention bypasses configured user/channel authorization
GHSA-j7qx-p75m-wp7gHigh· 7.5PraisonAI dynamic-context artifact tools read arbitrary host files outside artifact storage
PraisonAI dynamic-context artifact tools read arbitrary host files outside artifact storage
GHSA-fc26-m9pf-v56qHigh· 8.6PraisonAI LinearBot processes unsigned webhooks when LINEAR_WEBHOOK_SECRET is missing
PraisonAI LinearBot processes unsigned webhooks when LINEAR_WEBHOOK_SECRET is missing
GHSA-63v4-w882-g4x2High· 8.8PraisonAI: HTTPApproval dashboard renders tool arguments as raw HTML, allowing approval-page XSS to approve dangerous tools
PraisonAI: HTTPApproval dashboard renders tool arguments as raw HTML, allowing approval-page XSS to approve dangerous tools
GHSA-v847-hxxw-3pxgHigh· 7.8PraisonAI recipe.run_stream skips dangerous-tool policy enforcement
PraisonAI recipe.run_stream skips dangerous-tool policy enforcement
GHSA-w6h2-fr4q-xvxvHigh· 8.8PraisonAI: Compute-bridged file tools allow shell command injection
PraisonAI: Compute-bridged file tools allow shell command injection