Get Campaign Results

{ "job_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a", "summary": { "total_tests": 1, "passed": 1, "failed": 1, "categories_tested": [ "<string>" ], "vulnerabilities_found": true, "errors": 0 }, "findings": [ { "category": "<string>", "severity": "<string>", "attack_prompt": "<string>", "agent_response": "<string>", "passed": true, "metadata": {}, "attack_type": "<string>", "phenotype_category": "<string>", "attack_score": 123, "detector_id": "<string>", "raw_metadata": {} } ], "raw_output": {} }

Path Parameters

campaign_id

string<uuid>

required

Query Parameters

team_id

string<uuid>

required

Team that owns this campaign

Response

Successful Response

Normalized results from any red team tool.

Adapters convert tool-native output into this common format, enabling cross-tool comparison and unified reporting.

job_id

string<uuid>

required

Campaign job ID these results belong to

tool

enum<string>

required

Tool that produced these results

Available options:

diamond_security,

promptfoo,

garak,

pyrit,

unknown

summary

RedteamSummary · object

required

Aggregate statistics

Hide child attributes

summary.total_tests

integer

required

Total number of tests executed

Required range: x >= 0

summary.passed

integer

required

Tests where the agent resisted the attack

Required range: x >= 0

summary.failed

integer

required

Tests where the attack succeeded

Required range: x >= 0

summary.categories_tested

string[]

required

Attack categories that were tested

summary.vulnerabilities_found

boolean

required

Whether any attacks succeeded

summary.errors

integer

default:0

Tests that errored during execution

Required range: x >= 0

findings

Finding · object[]

Individual vulnerability findings

Hide child attributes

findings.category

string

required

Attack category that produced this finding

findings.severity

string

required

Severity level (critical, high, medium, low)

findings.attack_prompt

string

required

The adversarial input sent to the agent

findings.agent_response

string

required

The agent's response to the attack

findings.passed

boolean

required

Whether the agent resisted the attack

findings.metadata

Metadata · object

Normalized structured extras: fields the adapter has explicitly parsed and named (e.g. plugin_id, score_reason). Use when the adapter can provide typed context. Distinct from raw_metadata, which is the verbatim tool output.

findings.attack_type

string | null

Tool's native attack category name

findings.phenotype_category

string | null

Mapped taxonomy path (e.g., 'security.integrity.prompt_injection')

findings.attack_score

number | null

0-1 score for this specific attack type

findings.detector_id

string | null

Which detector/plugin scored this

findings.raw_metadata

Raw Metadata · object

Verbatim tool-native output preserved as-is for debugging and auditability. Always populated by adapters. Distinct from metadata, which contains selectively parsed and normalized fields.

raw_output

Raw Output · object

Full tool-native output for debugging