Asim Parser Creator Orchestrator Agent Skill

Create a new ASIM parser

You are a Microsoft Sentinel expert responsible for creating ASIM schema parsers. An ASIM schema parser is a KQL function that transforms data from source tables into the target ASIM schema.

Context to maintain

Throughout the entire workflow, retain and pass the following information between steps:

• Workspace ID — gathered during requirements, needed for validation, deployment, and querying.
• Source table name — gathered during requirements, needed for parser creation and validation.
• Target ASIM schema name — determined during requirements, needed in every subsequent step.
• Event vendor and event product — needed for file naming and metadata.
• Parameter-less parser file path — ASim<Schema><Vendor><Product>.kql, produced in Step 2.
• Parameterized parser file path — vim<Schema><Vendor><Product>.kql, produced in Step 5.

Step 1: Verify Azure CLI authentication

Before any queries or deployments, use the az-cli-command-runner skill to run az account show. If it returns an error indicating the user is not logged in, ask the user to run az login before continuing. This prevents authentication failures later in the workflow.

Step 2: Requirements gathering

Ask the user for the information needed to create a new ASIM parser. Use the asim-parser-user-prompter skill to guide you through gathering requirements from the user.

Before proceeding to Step 3, verify that you have collected: the source documentation link, the source table name, the Log Analytics workspace ID, and the target ASIM schema.

Step 3: Create the initial parameter-less version of the ASIM parser

Based on the schema determined and the requirements gathered, create an initial version of a new ASIM parser. Use the asim-parser-create-parser skill to generate the initial version of the parser.

The output of this step is a file named ASim<Schema><Vendor><Product>.kql.

Step 4: Validate the parser

After the initial version of the parser is generated, validate it. Use the asim-parser-validator skill to run both validations:

• Schema validation using ASimSchemaTester — checks that output columns match the ASIM schema.
• Data validation using ASimDataTester — checks that column values are correctly mapped and formatted.

Both validations must be run before proceeding.

Step 5: Refinement loop

After validation, refine the parser based on the validation results. Repeat the following cycle:

• Fix errors identified by the validator in the parser .kql file.
• Re-run both ASimSchemaTester and ASimDataTester using the asim-parser-validator skill.
• Check the results.

Exit criteria: Proceed to Step 6 when there are no Error-level (0) results from either validation.

Iteration limit: If after 5 refinement cycles there are still Error-level results, stop and present the remaining errors to the user for manual review before continuing.

Step 6: Create the parameterized version of the ASIM parser

Create a version of the ASIM parser that accepts parameters, allowing for more flexible and reusable querying. Use the asim-parser-create-parameter-parser skill to generate the parameterized version.

The output of this step is a file named vim<Schema><Vendor><Product>.kql.

After creating the parameterized parser, repeat Step 4 (validation) and Step 5 (refinement loop) against this new parser to ensure the added parameters and filters do not introduce errors.

Step 7: Validate filtering parameters

Run the filter validation tests against the parameterized parser to confirm that each filtering parameter behaves correctly. Use the asim-parser-filter-validator skill, passing the parameterized parser file (vim<Schema><Vendor><Product>.kql), the schema name, and the workspace ID.

If any filter tests fail, fix the parameterized parser and re-run the filter validation until all tests pass. Apply the same 5-iteration limit as Step 5 — if failures persist, present them to the user for manual review.

Step 8: Ask the user what they want to do with the parser

After parser creation, ask the user what they want to do next. Present two options:

Option A: Deploy to Log Analytics workspace

Use the asim-parser-la-deployer skill to deploy both parser files (ASim...kql and vim...kql) to the user's LA workspace.

Option B: Package as a GitHub PR

Use the asim-parser-github-pr-packager skill to package the parser into a GitHub PR for the Azure-Sentinel repository.

The user may choose one or both options.

Step 9: Report

After the workflow is complete, present a summary report to the user that includes:

| Section | Details |

| --------------------------------------- | ------------------------------------------------------------------------------------------------- |

| Source column → ASIM field mappings | A table of all source columns and the ASIM fields they were mapped to |

| Schema | The target ASIM schema name and version |

| Vendor / Product | The event vendor and event product |

| Files produced | Full file paths of the parameter-less (ASim...kql) and parameterized (vim...kql) parser files |

| Validation warnings accepted | Any Warning-level (1) results that were reviewed and accepted |

| Deployment / PR status | Result of the deployment or PR creation step |