Execution & Results

Complete reference for MCP tools that trigger test runs, poll for completion, and retrieve results including screenshots and step details.

Who is this for? SDETs, developers, and DevOps engineers integrating ContextQA with AI coding assistants (Claude, Cursor) or CI/CD pipelines.

These tools form the core execution loop: trigger a run, poll until it finishes, and retrieve the full evidence package.

---

execute_test_case

Triggers a single test case execution and returns a monitoring URL.

Category: Execution & Results Authentication required: Yes

Parameters

Name	Required	Type	Description
test_case_id	✅	integer	Numeric ID of the test case to run
run_in_portal	❌	boolean	Return a portal URL for live monitoring (default: true)
persona_id	❌	string	ID of a custom AI agent persona to use for this run
knowledge_id	❌	string	ID of a knowledge base to inject into this run

Returns

When run_in_portal is true: a portal URL string like https://app.contextqa.com/td/cases/18688/steps.

When false: a terminal command string for local execution.

Execution pattern

from time import sleep

# 1. Trigger
result = execute_test_case(test_case_id=18688)

# 2. Poll until terminal state
while True:
    status = get_execution_status(test_case_id=18688, number_of_executions=1)
    if "STATUS_COMPLETED" in status or "STATUS_FAILED" in status:
        break
    sleep(5)

# 3. Parse result_id from status string and retrieve results
results = get_test_case_results(result_id=47284)

Related Tools

get_execution_status • get_test_case_results • get_execution_step_details

---

get_execution_status

Polls for the current execution status of a test case. Call this repeatedly after execute_test_case until a terminal state is returned.

Category: Execution & Results Authentication required: Yes

Parameters

Name	Required	Type	Description
test_case_id	✅	integer	The test case whose most recent execution you want to check
number_of_executions	✅	integer	Total number of executions triggered since last check; typically 1

Returns

String containing the current status and result ID when complete. Possible statuses:

• Execution in progress — still running; poll again
• STATUS_COMPLETED, result_id: <id> — finished (check get_test_case_results for pass/fail)
• STATUS_FAILED, result_id: <id> — execution infrastructure failure

Notes

• Recommended polling interval: 5–10 seconds to avoid rate limiting.
• The result_id returned here is the ID to pass to get_test_case_results, get_execution_step_details, and other telemetry tools.

Related Tools

execute_test_case • get_test_case_results

---

get_test_case_results

Returns the complete result object for a test execution, including overall pass/fail, step count, duration, and all evidence URLs.

Category: Execution & Results Authentication required: Yes

Parameters

Name	Required	Type	Description
execution_id	❌	string	Execution ID (from execute_test_case return or portal URL)
result_id	❌	integer	Result ID (from get_execution_status return)

At least one of execution_id or result_id must be supplied.

Returns

JSON object with:

• result_id — unique result identifier
• status — PASSED or FAILED
• total_steps — total step count
• failed_steps — number of failed steps
• duration_ms — total execution time
• video_url — pre-signed S3 URL to the screen recording
• started_at / completed_at — ISO 8601 timestamps
• steps — array of step result summaries

Example

{
  "result_id": 47284
}

Related Tools

get_execution_step_details • get_test_step_results • get_root_cause

---

get_execution_step_details

Returns a human-readable step-by-step breakdown of a test execution with screenshots, failure reasons, and the overall verdict.

Category: Execution & Results Authentication required: Yes

Parameters

Name	Required	Type	Description
result_id	✅	integer	Result ID from get_execution_status or the portal URL

Returns

JSON object with:

• test_case_name — name of the test case
• overall_status — PASSED or FAILED
• total_steps — total step count
• failed_steps — number that failed
• screen_recording_url — video URL
• steps — array of per-step objects, each with:

  • step_number (1-based)
  • action — the step description
  • status — PASSED or FAILED
  • screenshot_url — URL of the step screenshot
  • failure_reason — AI explanation of what went wrong (if failed)

Example

{
  "result_id": 47284
}

Use case

This is the primary tool for a QA agent that wants to report test results to a developer. It provides all the evidence in a single structured call.

Related Tools

get_test_case_results • get_network_logs • get_console_logs • get_root_cause

---

fix_and_apply

Orchestrates the full failure-to-fix pipeline in a single call: fetches execution results, runs AI root cause analysis, queries the source repository for the responsible code, and returns a structured fix suggestion.

Category: Execution & Results Authentication required: Yes

Parameters

Name	Required	Type	Description
execution_id	✅	string	ID of the failed execution
repo_url	✅	string	URL of the source code repository to query (GitHub, GitLab, Bitbucket)

Returns

JSON object with:

• execution_id — echoed back for reference
• test_case_name — which test failed
• failure_description — plain English description of the failure
• root_cause_analysis — AI analysis of what caused it
• code_context — relevant code snippets from the repository
• next_steps — specific actions recommended (file to edit, line to change)

Use case

Call this after a test failure when you have access to the source repository. The tool automatically:

1. Fetches the execution details
2. Identifies the failed step and its screenshot
3. Sends evidence to the root cause analysis engine
4. Searches the repository for code related to the failed feature
5. Returns a correlated analysis with actionable fix suggestions

Example

{
  "execution_id": "exec_abc123",
  "repo_url": "https://github.com/myorg/myapp"
}

Related Tools

get_root_cause • investigate_failure • get_execution_step_details