Cxas Self Steering Agent
Builds and refactors self-steering CXAS (CES, GECX) conversational agents for chat or voice, where the model runs the whole conversation like a contact center agent: one root agent, Jinja progressive disclosure (playbooks and tasks loaded through open_playbook and load_task, dynamic tool hiding) and deterministic actions (greeting, verbatim text, goodbye and live agent handoff through set_next_agent_actions). Ships a working sample agent to start from. Use when creating a new CXAS agent, adding playbooks, tasks or tools to an agent built this way, replacing a steering agent with sub-agents or
Install
npx skills add GoogleCloudPlatform/cxas-scrapi --skill cxas-self-steering-agentgit clone --depth 1 https://github.com/GoogleCloudPlatform/cxas-scrapi.git /tmp/cxas-scrapi
mkdir -p ~/.claude/skills
cp -r /tmp/cxas-scrapi/.agents/skills/cxas-self-steering-agent ~/.claude/skills/cxas-self-steering-agentPersonal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.
From the SKILL.md
# CXAS self-steering agent This skill builds self-steering CXAS agents. Treat the model like a capable person on a support team: give it the power to run the whole conversation and the awareness to know where it is and why, and set up its job so it can do both. There's no steering agent and no code making decisions behind the model's back; the agent steers itself. Progressive disclosure and deterministic actions are the two main techniques, built from a Jinja instruction template, Python tools, session state and one small callback on top of the loop CXAS runs. The skill bundles a working sample agent ([resources/sample-agent](resources/sample-agent)) and these scripts: - `scripts/scaffold_from_sample.py` copies the sample into a new app directory. - `scripts/check_tool_visibility.py` renders the instruction for every playbook and task, lists the tools the model can see, and catches mistakes `cxas lint` misses. - `scripts/sync_tool_descriptions.py` copies tool docstrings into the tool JSON files. - `scripts/run_locally.py` runs tools and the callback on your machine. - `scripts/run_conversation.py` runs a test conversation against the deployed app. Read [references/guide.md](references/guide.md) before designing or changing an agent. It has the reasoning, the code patterns, and the platform gotchas. ## Before you start
Details
- Skill name
cxas-self-steering-agent- Repository
- GoogleCloudPlatform/cxas-scrapi
- Path
.agents/skills/cxas-self-steering-agent/SKILL.md- License
- Apache-2.0
- Repo updated
- 17 November 2026
FAQ
What does the Cxas Self Steering Agent skill do?
Builds and refactors self-steering CXAS (CES, GECX) conversational agents for chat or voice, where the model runs the whole conversation like a contact center agent: one root agent, Jinja progressive disclosure (playbooks and tasks loaded through open_playbook and load_task, dynamic tool hiding) and deterministic actions (greeting, verbatim text, goodbye and live agent handoff through set_next_agent_actions). Ships a working sample agent to start from. Use when creating a new CXAS agent, adding playbooks, tasks or tools to an agent built this way, replacing a steering agent with sub-agents or
How do I install the Cxas Self Steering Agent skill?
skills CLI (Claude Code, Cursor, Codex and more): npx skills add GoogleCloudPlatform/cxas-scrapi --skill cxas-self-steering-agent. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.
Is Cxas Self Steering Agent popular?
It lives in GoogleCloudPlatform/cxas-scrapi, which has 117 GitHub stars and 117 forks.
Related skills
Adk Framework Ingestor
117GoogleCloudPlatform/cxas-scrapi
Parsing and ingestion directives for the Python-based Agent Development Kit (ADK) conversational agent framework.
Cxas Agent Foundry
117GoogleCloudPlatform/cxas-scrapi
End-to-end GECX/CXAS/CES conversational agent lifecycle -- build agents from requirements (PRD-to-agent), create and run evals (goldens, simulations, tool tests, callback tests), debug failures, and iterate to production quality. Use this skill whenever the user mentions GECX, CXAS, CES, SCRAPI, conversational agents, voice agents, audio agents, agent evals, pushing/pulling/linting agents, or agent instructions/callbacks/tools on the Google Customer Engagement Suite platform.
Cxas Autolabel Rules
117GoogleCloudPlatform/cxas-scrapi
Author, validate, and manage Contact Center AI (CCAI) Insights Autolabeling Rules. Use when users want to translate business labeling logic into CEL condition expressions, maintain declarative autolabel_rules.yaml configurations, or synchronize rules to GCP projects.
Cxas Sim Eval
117GoogleCloudPlatform/cxas-scrapi
Converts CXAS golden evaluations to SCRAPI SimulationEvals test cases. Use when generating high-level, goal-oriented test cases from turn-by-turn evaluation JSONs, and when enriching test expectations with inferred tool calls.
Cxas Protocol Robust Extraction
117GoogleCloudPlatform/cxas-scrapi
A robust methodology for LLM-based requirements gathering and high-fidelity artifact generation. Employs 'Divide, Conquer, and Verify' tactics using specialized subagents, iterative exhaustion loops, and batched execution to ensure zero data loss.
Doc Sync Tool
193MavCui/Agent-Skills-Hunter
自动同步项目中的 Agents.md、claude.md 和 gemini.md 文件,保持内容一致性。支持自动监听和手动触发。
Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.