Can My Agent Uselookup · compare · cite

Aug 28, 2026 · Catalog foundation is up

Product feature · security-privacy

Secrets managementCommon product term

Store and provide credentials through a documented secrets feature.

Terminology basis: Common product term. GitHub — Secrets reference

Secrets management: 4 supported, 0 partial, 0 unsupported, 27 unreviewed across 31 cataloged products.

Markdown · JSON

Explore this familyMore in Security and privacy7 capabilities

Current evidence by product

Can my agent use Secrets management?

Read across for the answer. 4 of 31 current product columns have reviewed evidence; unreviewed does not mean unsupported.

Known issues (0)Report an issue
  • Supported4
  • Partial0
  • Unsupported0
  • Unknown27
  • Not applicable0

Web

9 products

Desktop

13 products

CLI

9 products

Unknown means no public evidence has been reviewed for that product and capability. It does not mean unsupported.

How statuses are assigned

Definition and scope

What this capability means

This row asks how a harness handles credentials used by tools, connectors, repositories, terminals, MCP servers, and sub-agents. Plain environment variables or pasted API keys are not equivalent to a managed secret that can be injected without revealing its value to model context or logs.

Evidence should record storage boundary, encryption, user and workspace scope, tool allowlists, just-in-time grants, masking, rotation, revocation, export prevention, prompt-injection defenses, child-agent inheritance, and audit events. Redaction is best-effort unless the documented mechanism prevents the model from receiving the value in the first place.

Traceable compatibility

Assertion ledger

Documentation evidence only. No runtime conformance test is implied.

Cursordesktop · current
Supported
Target
2026-08-28 Cursor Background Agent secrets documentation observation · hosted-observation
Environment
hosted-default
Observed
2026-08-28
  • runtimedevelopment-environment secrets are stored encrypted at rest using KMS and injected into the remote Background Agent environment
  • policyper-tool access controls, log masking, rotation, and prevention of model or shell access to injected values are not established by the reviewed page
Evidence
Copilotweb · current
Supported
Target
2026-08-28 Copilot cloud-agent secrets documentation observation · hosted-observation
Environment
enterprise-managed
Observed
2026-08-28
  • policydedicated Agents secrets can be scoped to a repository or organization, with all, private, or selected repository access for organization values
  • runtimevalues are injected as environment variables and masked in session logs; COPILOT_MCP_-prefixed values are exposed only to MCP servers
  • policythe cloud agent cannot access Actions, Codespaces, or Dependabot secrets and variables
Evidence
Devinweb · current
Supported
Target
2026-08-28 Devin secrets documentation observation · hosted-observation
Environment
hosted-default
Observed
2026-08-28
  • policyencrypted-at-rest secrets support organization-wide, personal, repository, and session scope; global values are usable by all organization members but viewable and editable only by admins
  • formatsupported credential forms include raw values, site cookies, and TOTP material
  • runtimeconfigured values are injected as environment variables for sessions created after configuration, so Devin and executed code can access them
Evidence
Replit Agentweb · current
Supported
Target
2026-08-28 Replit Secrets documentation observation · hosted-observation
Environment
hosted-default
Observed
2026-08-28
  • policyapp-scoped and account-scoped secrets are encrypted with AES-256 at rest and TLS in transit and can be linked to selected Replit Apps
  • runtimevalues are supplied to the app and Agent execution environment as environment variables; static deployments are excluded
  • policycollaborators and non-owner organization members may be unable to reveal a value in the UI yet can still access it by printing the environment variable
Evidence
  1. 1. Evidence checked 2026-08-28: Cursor Background Agents accept development-environment secrets, store them encrypted at rest with KMS, and inject them into the remote agent environment. The reviewed page does not document per-tool allowlists, masking behavior, or a guarantee that the model cannot read an injected value.
  2. 2. Evidence checked 2026-08-28: GitHub Copilot cloud agent has dedicated Agents secrets at repository and organization scope, repository allowlists for organization secrets, masked session logs, and an MCP-only prefix that withholds designated values from the general agent environment. It cannot access Actions, Codespaces, or Dependabot secrets.
  3. 3. Evidence checked 2026-08-28: Devin encrypts global secrets at rest and supports organization, personal, repository, and session scopes, including raw credentials, site cookies, and TOTP material. Organization members can use global secrets while only admins can view or edit them; environment injection means Devin or executed code can access the value.
  4. 4. Evidence checked 2026-08-28: Replit's Secrets tool encrypts app and account secrets with AES-256 at rest and TLS in transit and exposes linked values as environment variables to the app and Agent execution environment. Collaborators and non-owner organization members may still recover values by printing environment variables, which is an important disclosure boundary.