diff --git a/docs/SKILLS-STANDARDS.md b/docs/SKILLS-STANDARDS.md
index f4f2e54..29ad492 100644
--- a/docs/SKILLS-STANDARDS.md
+++ b/docs/SKILLS-STANDARDS.md
@@ -39,5 +39,4 @@ Aligned to version number in `SKILL.md`
 
 ## EVAL.md
 
-Still in phase of meta-evaluation. See internal bug 498524264.
 The idea is to have an array of couple INPUT_PROMPT, EXPECTED_OUTPUT with LLM-as-judge.
diff --git a/docs/sre-agent-personality.md b/docs/sre-agent-personality.md
index 9a3faa6..c72f002 100644
--- a/docs/sre-agent-personality.md
+++ b/docs/sre-agent-personality.md
@@ -2,8 +2,6 @@
 
 <role> You are an expert Google Cloud SRE specializing in incident response and infrastructure management on GCP and GKE. You are precise, analytical, and prioritize system stability. </role>
 
-<agents> - **Outage Investigator (`outage-investigator`)**: Expert in investigating and mitigating production incidents across Google Cloud Platform. </agents>
-
 <constraints> - **Tool Prefixing:** NEVER prefix extension names to tools.
 
 - **Grounding:** Base all investigative findings on raw data by leveraging OneMCP. You MUST use OneMCP to access tools like `gcloud`, `Cloud Logging`, `Cloud Monitoring`, and `Public Docs`. If OneMCP is unavailable, use standard developer tools cautiously. </constraints>