Composio powers 1000+ toolkits, tool search, context management, authentication, and a sandboxed workbench to help you build AI agents that turn intent into action.
docs(changelog): drop self-confessing 'previously plaintext' sentence
Two reasons to kill it: - A security changelog doesn't need to catalog our prior bug state. The preceding sentence already tells the customer what to do if they depend on the value; reading "these fields were previously returned in plaintext" adds no action they can take, just public self-confession. - Customers who were reading the value already know they were reading it; customers who weren't don't benefit from the footnote. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Anshu Garg committed
Apr 22, 2026 at 12:21 UTC
2d7bb6305a13fb275e79b855f2831bafb4e4114c
1 file changed
+1
-1
docs/content/changelog/04-22-26-composio-managed-credentials-masking.mdx
+1
-1
@@ -73,7 +73,7 @@ If you were reading Composio-managed access or refresh tokens from these respons
73
- **Composio-managed auth configs** — every credential field is redacted.
74
- **Customer-owned auth configs** — public OAuth identifiers (`client_id`, `consumer_key`) still pass through so you can debug OAuth flows. Every other credential field is obfuscated.
75
76
-If your code reads a secret — for example `client_secret` or `developer_token` — out of these responses, you will need to keep that value alongside the code that created the auth config (typically in your own secret manager at upload time). These fields were previously returned in plaintext on some code paths and are now obfuscated.
76
+If your code reads a secret — for example `client_secret` or `developer_token` — out of these responses, you will need to keep that value alongside the code that created the auth config (typically in your own secret manager at upload time).
77
78
### Why this change
79