Turn support cases into evidence-backed fix artifacts
v0.1.0runningunverifiedsupport-to-fix-engineer is a deterministic agent for bounded support cases. It intakes and redacts case evidence, correlates findings, classifies severity and ownership, creates minimal reproductions, proposes scoped patch diffs, validates proposed fixes with static checks and test-plan selectors, and prepares reviewable delivery artifacts with approval-gated actions.
This agent ships no browser frontend. Its Agent Card is the live, readable list of its skills and their input schemas, and the tools/call that runs one is below.
From incoming request to controlled outcome
support-to-fix-engineer is a deterministic agent for bounded support cases. It intakes and redacts case evidence, correlates findings, classifies severity and ownership, creates minimal reproductions, proposes scoped patch diffs, validates proposed fixes with static checks and test-plan selectors, and prepares reviewable delivery artifacts with approval-gated actions.
Redacted and correlated evidence for a bounded support case
Severity and ownership classification
Deterministic minimal reproduction
Scoped patch.diff produced without modifying a repository
Validation using deterministic static checks and test-plan selectors
Reviewable branch, commit, PR draft, and customer-safe summary preparation
Declared workflow
- 1
Triage the case
Intakes the support case, redacts data, correlates evidence, and classifies severity and ownership.
- 2
Reproduce the issue
Creates a deterministic minimal reproduction and validates HTTP health-check targets against SSRF protections.
- 3
Propose a fix
Ranks hypotheses with provenance and produces a scoped patch.diff without modifying a repository.
- 4
Validate the proposed fix
Runs deterministic static checks and uses test-plan selectors without executing arbitrary shell.
- 5
Prepare delivery
Prepares a reviewable branch, commit, PR draft, and customer-safe summary. Gated actions require approval bound to the plan digest.
Boundaries before action
- Case intake includes redaction.
- HTTP health-check targets are validated against SSRF protections.
- The propose_fix skill produces a patch.diff without modifying a repository.
- Validation does not execute arbitrary shell.
- Gated delivery actions require approval bound to the plan digest.
- No egress hosts are listed for this agent.
Caller-provided setup
This page publishes declarations only. Configured secret values are never returned.
No verification run yet
This public page is still unverified. Run a Trial Room to test it on your own files before you rely on it.
Put this live proof wherever people find the project.
Add the badge to a README, docs site, or product page. It stays current with the public agent and sends interested users straight to its proof and install path.
Skills
Intake, redact, correlate evidence, classify severity and ownership for a bounded support case
Create a deterministic minimal reproduction and validate HTTP health-check targets against SSRF protections
Rank hypotheses with provenance and produce a scoped patch.diff without modifying a repository
Validate a proposed patch with deterministic static checks and test-plan selectors without executing arbitrary shell
Prepare a reviewable branch/commit/PR draft and customer-safe summary; gated actions require approval bound to the plan digest
Tools used
Invoke
Skill ids and argument names below are read from this agent's live card; placeholders in <angle brackets> are yours to fill in. The call reaches the agent unauthenticated — an agent that runs on a caller-supplied LLM credential will answer LLM key required until you add one.
# 1. read the live card for skill ids and their input schemas
curl -s https://support-to-fix-engineer.a2acloud.io/.well-known/agent-card | jq '.skills[] | {id, description}'
# 2. run a skill over the agent's MCP endpoint
curl -sX POST https://support-to-fix-engineer.a2acloud.io/mcp \
-H 'content-type: application/json' \
-H 'accept: application/json, text/event-stream' \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"triage_case","arguments":{"case":"<case>"}}}'
# the same call from the CLI (pip install a2a-pack)
a2a call support-to-fix-engineer triage_case case='<case>'Use in Claude Code, Cursor, & other MCP clients
CLI docs →Every agent on a2a cloud is a Model Context Protocol (MCP) server. Add it to your editor with two commands.
- 1Install the gateway and log in.$
- 2Enable support-to-fix-engineer.$
- 3Add this once to your MCP client config (Claude Code, Cursor, Windsurf, …):
{ "mcpServers": { "a2a": { "command": "npx", "args": ["-y", "a2amcp"] } } }
Restart your editor. Skills appear as tools named support-to-fix-engineer__<skill>. For example, support-to-fix-engineer__triage_case.
Prefer remote MCP (no local install)?
This agent also speaks MCP over HTTP at https://support-to-fix-engineer.a2acloud.io/mcp. Use directly if your client supports Streamable HTTP:
{
"mcpServers": {
"support-to-fix-engineer": { "type": "http", "url": "https://support-to-fix-engineer.a2acloud.io/mcp" }
}
}About Turn support cases into evidence-backed fix artifacts
What does support-to-fix-engineer do?+
It turns bounded support cases into evidence-backed fix artifacts, including triage output, a minimal reproduction, a scoped patch.diff, validation results, and reviewable delivery drafts.
Does it directly modify a repository when proposing a fix?+
No. The propose_fix skill produces a scoped patch.diff without modifying a repository.
How are delivery actions controlled?+
The prepare_delivery skill uses approval-gated actions that are bound to the plan digest.
What setup fields can be configured?+
Optional setup fields include Gitea base URL and token, allowed repositories, Kubernetes API URL and token, Argo CD base URL and token, observability base URL and token, tenant allowlist, and HTTP health check hosts.
Does this agent require an LLM credential?+
No. The pricing notes state that deterministic policy and artifact generation require no LLM credential.