Composio powers 1000+ toolkits, tool search, context management, authentication, and a sandboxed workbench to help you build AI agents that turn intent into action.
| 1 | # Security Policy |
| 2 | |
| 3 | ## Reporting a Vulnerability |
| 4 | |
| 5 | We take the security of Composio seriously. If you believe you have found a security vulnerability, please report it to us through GitHub Security Advisories or email us at **security@composio.dev** |
| 6 | |
| 7 | ### How to Report |
| 8 | |
| 9 | **Please do not report security vulnerabilities through public GitHub issues.** |
| 10 | |
| 11 | Instead, please report them using one of the following methods: |
| 12 | |
| 13 | - **GitHub Security Advisory (Preferred)**: Report a vulnerability directly through GitHub by visiting: [https://github.com/composiohq/composio/security/advisories/new](https://github.com/composiohq/composio/security/advisories/new) |
| 14 | - **Email**: If you prefer not to use GitHub Security Advisories, you can email security concerns to **security@composio.dev** |
| 15 | |
| 16 | ### What to Include |
| 17 | |
| 18 | Please include as much of the following information as possible: |
| 19 | |
| 20 | - Type of vulnerability |
| 21 | - Full paths of source file(s) related to the vulnerability |
| 22 | - Location of the affected source code (tag/branch/commit or direct URL) |
| 23 | - Step-by-step instructions to reproduce the issue |
| 24 | - Proof-of-concept or exploit code (if possible) |
| 25 | - Impact of the vulnerability, including how an attacker might exploit it |
| 26 | |
| 27 | ### Response Timeline |
| 28 | |
| 29 | - We will acknowledge your report within **48 hours** |
| 30 | - We will provide a more detailed response within **7 days** indicating the next steps |
| 31 | - We will keep you informed of the progress toward resolving the issue |
| 32 | - We may ask for additional information or guidance |
| 33 | |
| 34 | ## Ground Rules |
| 35 | |
| 36 | ### Do: |
| 37 | |
| 38 | - Abide by these Program Terms |
| 39 | - Be patient and make a good faith effort to provide clarifications to any questions we may have about your submission |
| 40 | - Be respectful when interacting with our team, and our team will do the same |
| 41 | - Perform testing only using accounts that are your own personal/test accounts |
| 42 | - Exercise caution when testing to avoid negative impact to data or services |
| 43 | - Respect privacy and make a good faith effort not to change or destroy Composio or personal data |
| 44 | - Stop whenever you are unsure if your test case may cause, or have caused, destructive data or systems damage; report your initial finding(s) and request authorization to continue testing |
| 45 | |
| 46 | ### Do NOT: |
| 47 | |
| 48 | - Leave any system in a more vulnerable state than you found it |
| 49 | - Use or interact with accounts you do not own |
| 50 | - Brute force credentials or guess credentials to gain access to systems or accounts |
| 51 | - Change passwords of any account that is not yours or that you do not have explicit permission to change |
| 52 | - Perform denial of service (DoS) attacks or related tests that would cause availability interruptions or degradation of our services |
| 53 | - Publicly disclose a vulnerability submission without our explicit review and consent |
| 54 | - Engage in any form of social engineering of Composio employees, customers, or partners |
| 55 | - Engage or target any specific Composio employees, customers, or partners during your testing |
| 56 | - Access, extract, or download personal or business information beyond that which is minimally necessary for your Proof-of-Concept purposes |
| 57 | - Do anything that would cause destruction of Composio data or systems |
| 58 | - Test against production systems without permission |
| 59 | - Access or modify user data |
| 60 | - Exploit vulnerabilities beyond what is needed for a proof of concept |
| 61 | |
| 62 | ## Out-of-Scope Vulnerabilities |
| 63 | |
| 64 | Certain vulnerabilities are considered out-of-scope for the Bug Bounty Program. Those out-of-scope vulnerabilities include, but are not limited to: |
| 65 | |
| 66 | - Vulnerabilities not involving product or coding flaws, but solely relying upon possession of stolen or compromised credentials |
| 67 | - Vulnerabilities dependent on Phishing in a DNS domain that is not in one of our primary service domains |
| 68 | - Most vulnerabilities that rely on a runtime context within a sandbox, lab, staging, testing or non-production environment |
| 69 | - Vulnerabilities involving stolen or compromised credentials |
| 70 | - Open redirect resulting in a low security impact |
| 71 | - Credential stuffing or physical access to a device |
| 72 | - Any vulnerabilities requiring significant and unlikely interaction by the victim |
| 73 | - Man-in-the-Middle attacks except in mobile applications |
| 74 | - Account enumeration with a pre-defined and known list of UUIDs |
| 75 | - Invite/Promo code enumeration |
| 76 | - Ability to send push notifications/SMS messages/emails without the ability to change content |
| 77 | - Information disclosures related to existence of accounts |
| 78 | - Reports stating that a particular software component is vulnerable without an accompanying proof-of-concept |
| 79 | - Vulnerabilities only affecting users using outdated, unpatched, or unsupported browsers or software |
| 80 | - Stack traces, path disclosure, and directory listings |
| 81 | - CSV injection vulnerabilities |
| 82 | - Best practices concerns without a demonstrable information assurance issue and proof-of-concept |
| 83 | - Ability to take over social media pages |
| 84 | - Negligible security severity |
| 85 | - Speculative reports about theoretical damage – please always provide a proof-of-concept |
| 86 | - Self-XSS or similar vulnerabilities that cannot be used to exploit other users |
| 87 | - Vulnerabilities reported by automated scanning tools without additional analysis |
| 88 | - Distributed or denial of service attacks (DDoS/DoS) and/or reports on rate limiting issues |
| 89 | - Content injection or content spoofing issues |
| 90 | - Cross-site Request Forgery (CSRF) with minimal security implications |
| 91 | - Missing cookie flags on non-authentication cookies |
| 92 | - Submissions that require physical access to a victim's computer/device |
| 93 | - SSL/TLS protocol scan reports |
| 94 | - Banner grabbing issues |
| 95 | - Open ports or services without accompanying proof-of-concept |
| 96 | - Physical or social engineering attempts |
| 97 | - Exposed login panels without proof-of-concept |
| 98 | - Dangling IPs |
| 99 | - Subdomain takeovers without proof |
| 100 | - Reports on third-party products, services, or applications not owned by Composio |
| 101 | - Out-of-scope domains |
| 102 | - Leaked API Key or credentials having no impact on Composio assets |
| 103 | |
| 104 | ## CVEs and 0-Days |
| 105 | |
| 106 | ### CVE Age Requirement |
| 107 | |
| 108 | We will only accept CVEs that meet all of the criteria below: |
| 109 | |
| 110 | - Public for at least **60 days** |
| 111 | - Demonstrated impact on our environment (not theoretical) |
| 112 | - Reproducible with clear steps or proof-of-impact |
| 113 | |
| 114 | **CVE reports that do not meet the 60-day requirement are out of scope.** |
| 115 | |
| 116 | ## Supported Versions |
| 117 | |
| 118 | We release patches for security vulnerabilities. Please ensure you are using the latest version of Composio. |
| 119 | |
| 120 | ## Disclosure Policy |
| 121 | |
| 122 | - We follow coordinated disclosure practices |
| 123 | - Security advisories will be published after a fix is available |
| 124 | - We appreciate responsible disclosure and will acknowledge reporters in the advisory (unless you prefer to remain anonymous) |
| 125 | |
| 126 | --- |
| 127 | |
| 128 | **Thank you for helping keep Composio and our users safe!** |
| 129 | |
| 130 | We follow responsible disclosure practices and work with researchers to ensure vulnerabilities are properly addressed before public disclosure. |