Skip to content

Commit 182d7c2

Browse files
committed
Add environment yamls
1 parent ee2880d commit 182d7c2

File tree

4 files changed

+920
-0
lines changed

4 files changed

+920
-0
lines changed

tools/src/test/resources/environment/BUILD.bazel

Whitespace-only changes.
Lines changed: 105 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,105 @@
1+
# Copyright 2026 Google LLC
2+
#
3+
# Licensed under the Apache License, Version 2.0 (the "License");
4+
# you may not use this file except in compliance with the License.
5+
# You may obtain a copy of the License at
6+
#
7+
# https://www.apache.org/licenses/LICENSE-2.0
8+
#
9+
# Unless required by applicable law or agreed to in writing, software
10+
# distributed under the License is distributed on an "AS IS" BASIS,
11+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12+
# See the License for the specific language governing permissions and
13+
# limitations under the License.
14+
15+
# Inherits from the common_agent_env.yaml and extends it with additional
16+
# variables for tool call support.
17+
18+
name: "ai.agent_env"
19+
20+
variables:
21+
- name: "agent.name"
22+
type_name: "string"
23+
description: |
24+
The name of the agent.
25+
- name: "agent.description"
26+
type_name: "string"
27+
description: |
28+
A description of the agent.
29+
- name: "agent.auth"
30+
type_name: "cel.expr.ai.AgentAuth"
31+
description: |
32+
The authentication information for the agent.
33+
Valid agent auth properties include:
34+
- agent.auth.principal: The principal associated with the agent.
35+
- agent.auth.claims: The claims asserted by the agent.
36+
- agent.auth.oauth_scopes: The scopes required by the agent.
37+
- name: "agent.context"
38+
type_name: "cel.expr.ai.AgentContext"
39+
description: |
40+
The context of the agent which includes the user input, history, and other relevant data.
41+
Valid agent context properties include:
42+
- agent.context.prompt: string-typed representation of the prompt provided to the LLM,
43+
combining system instructions, context, and user input.
44+
- agent.context.trust.level: string-typed trust level of the agent context.
45+
- agent.context.trust.findings: list of cel.expr.ai.Finding values which which contribute
46+
to the trust level.
47+
- agent.context.sources: list of sources referenced in the agent context where each source
48+
contains a 'name' and 'value'. The name might be a resource name, the value could either
49+
be a string, base64-encoded bytes, or uri if present.
50+
- name: "agent.model"
51+
type_name: "cel.expr.ai.Model"
52+
description: |
53+
The model used by the agent.
54+
Valid agent model properties include:
55+
- agent.model.name: The name of the model.
56+
- name: "agent.provider"
57+
type_name: "cel.expr.ai.AgentProvider"
58+
description: |
59+
The provider of the model used by the agent.
60+
Valid agent provider properties include:
61+
- agent.provider.url: The url where the agent can be found, either a service endpoint or
62+
an agent card uri: "https://<agent-card-domain>/<agent-id>/.well-known/agent-card.json".
63+
- agent.provider.organization: The organization which maintains the agent.
64+
65+
- name: "agent.input"
66+
type_name: "cel.expr.ai.AgentMessage"
67+
description: |
68+
The input to the agent, represented as a cel.expr.ai.AgentMessage.
69+
Valid agent message properties include:
70+
- agent.input.role: string role of the message, either 'user' or 'agent'.
71+
- agent.input.metadata: dynamic metadata associated with the input.
72+
- agent.input.time: The timestamp of the message.
73+
- agent.input.parts: list of message parts provided in the input.
74+
Each message part is a cel.expr.ai.AgentMessage.Part.
75+
76+
To inspect the message parts use helper methods:
77+
- agent.input.safetyFindings(<label_name>): Returns the safety findings associated with the
78+
parts with the given label name.
79+
- agent.input.sensitivityFindings(<label_name>): Returns the sensitivity findings
80+
associated with the label for all parts of the message.
81+
- agent.input.threatFindings(): Returns the threat findings associated with the
82+
parts of the message.
83+
parts with the given label name.
84+
85+
- name: "agent.output"
86+
type_name: "cel.expr.ai.AgentMessage"
87+
description: |
88+
The output to respond with from the agent, represented as a cel.expr.ai.AgentMessage.
89+
Valid agent message properties include:
90+
- agent.output.role: string role of the message, either 'user' or 'agent'.
91+
- agent.output.metadata: dynamic metadata associated with the input.
92+
- agent.output.time: The timestamp of the message.
93+
- agent.output.parts: list of message parts provided in the input.
94+
Each message part is a cel.expr.ai.AgentMessage.Part.
95+
96+
To inpect the findings in the message parts use helper methods:
97+
- agent.output.safetyFindings(<label_name>): Returns the safety findings associated with the
98+
parts with the given label name.
99+
- agent.output.sensitivityFindings(<label_name>): Returns the sensitivity findings
100+
associated with the label for all parts of the message.
101+
- agent.output.threatFindings(): Returns the threat findings associated with the
102+
parts of the message.
103+
104+
- name: "agent.history"
105+
type_name: "cel.expr.ai.AgentMessageSet"

0 commit comments

Comments
 (0)