trestle-control-implementation

name	trestle-control-implementation
description	Knowledge about writing control implementation responses, rules, parameters, component-level responses, inheritance, and leveraged SSPs in Compliance Trestle. Use when users ask about writing control responses, implementation status, rules, parameters, component definitions, SSP implementation details, or compliance documentation content.
allowed-tools	Bash, Read, Glob, Grep, Write, Edit

Control Implementation in Trestle

Writing Control Responses

SSP Control Markdown Structure

Each control markdown file in an SSP has this structure:

---
x-trestle-set-params:
  ac-1_prm_1:
    values:
      - organization-defined value
    ssp-values:
      - actual SSP value
x-trestle-comp-def-rules:
  My Component:
    - name: rule-ac-1
      description: Ensure access control policy exists
x-trestle-comp-def-rules-param-vals:
  My Component:
    - name: rule-param-1
      values:
        - default value
      ssp-values:
        - SSP override value
---

# ac-1 - [Access Control] Policy and Procedures

## Control Statement
[Control statement text with {{ insert: param, ac-1_prm_1 }} placeholders]

## Control guidance
[Guidance text]

______________________________________________________________________

## Implementation for part a.

### My Component

<!-- Add control implementation description here for part a. -->
Implementation prose for component addressing part a.

#### Rules:
  - rule-ac-1

#### Implementation Status: implemented

### This System

<!-- Add control implementation description here for part a. -->
Implementation prose for the overall system addressing part a.

#### Implementation Status: partial

Component-Level Responses

Each control statement part gets responses from each component:

1. Named Components (from component-definitions): Have rules, parameters, status
2. This System component: Overall system-level response (always present)

First run without component definitions: If you run ssp-generate without --compdefs (or before any component-definitions exist in the workspace), the generated markdown will only contain This System sections. Named component sections such as Identity Provider in the examples below only appear after component definitions are loaded and passed via --compdefs. This is expected, not an error.

Adding Implementation Prose

Replace the HTML comment placeholders with actual implementation text:

### My Component
This component implements access control policy by...

Rules and Parameters

Rules

Rules come from component definitions and are read-only in SSP markdown:

x-trestle-comp-def-rules:
  Component Name:
    - name: rule-id
      description: What the rule checks

Rule Parameters

Parameters for rules can have values overridden in SSP:

x-trestle-comp-def-rules-param-vals:
  Component Name:
    - name: param-name
      values:
        - component default
      ssp-values:
        - SSP override

Implementation Status

Set per-component using these values:

Status	Meaning
implemented	Fully implemented
partial	Partially implemented
planned	Implementation planned
alternative	Alternative implementation exists
not-applicable	Not applicable to this component

In markdown: #### Implementation Status: implemented

Inheritance and Leveraged SSPs

Leveraged SSPs

When a system inherits controls from a provider:

## This System: Component Name

### Provided Statement Description
Description of what the provider system provides.

### Responsibility Statement Description
Description of customer responsibilities.

### Satisfied Statement Description
How the inheriting system satisfies its responsibilities.

Control Origination

Tracks where control implementation comes from:

• organization - Organization-wide policy/procedure
• system-specific - Specific to this system
• customer-configured - Customer configures provider capability
• customer-provided - Customer provides the implementation
• inherited - Inherited from provider system

Profile-Level Control Additions

Profiles can add sections to controls:

x-trestle-sections:
  guidance: Guidance
  implgdn: Implementation Guidance
  expevid: Expected Evidence

These sections appear in the markdown and can be edited:

## Implementation Guidance
Organization-specific implementation guidance here.

## Expected Evidence
Evidence required to demonstrate implementation.

Best Practices

1. Be specific: Reference actual system components, tools, and processes
2. Address each part: Ensure every statement part has a response
3. Set parameters: Replace <REPLACE_ME> placeholders with actual values
4. Track status honestly: Use partial or planned when appropriate
5. Document inheritance: Clearly describe what is inherited vs. locally implemented
6. Use component definitions: Define reusable compliance content in component-definitions
7. Leverage CI/CD: Use assemble in pipelines to validate changes automatically

Worked Example: Writing AC-2 Account Management

This example assumes a component definition providing an Identity Provider component was passed to ssp-generate via --compdefs. Without it, only the This System sections appear.

Before: Generated Markdown (Unfilled)

---
x-trestle-set-params:
  ac-2_prm_1:
    values:
      - organization-defined account types
    ssp-values:
---

# ac-2 - Account Management

## Control Statement

...

______________________________________________________________________

## Implementation for part a.

### Identity Provider

<!-- Add control implementation description here for part a. -->

#### Implementation Status: planned

### This System

<!-- Add control implementation description here for part a. -->

#### Implementation Status: planned

After: Filled-In Implementation

---
x-trestle-set-params:
  ac-2_prm_1:
    values:
      - organization-defined account types
    ssp-values:
      - privileged, non-privileged, system, service, and temporary accounts
---

# ac-2 - Account Management

## Control Statement

...

______________________________________________________________________

## Implementation for part a.

### Identity Provider

The Identity Provider (IdP) manages all user account types including privileged,
non-privileged, and service accounts. Account provisioning is handled through
the IdP's administrative console with SCIM integration to downstream systems.

#### Rules:
  - rule-ac-2-account-types

#### Implementation Status: implemented

### This System

The system integrates with the organization's Identity Provider for account
lifecycle management. Local service accounts are managed through
infrastructure-as-code templates with mandatory approval workflows.

#### Implementation Status: implemented

## Implementation for part b.

### Identity Provider

Account managers are designated through the IdP's delegation model. Each
organizational unit has a primary and backup account manager assigned in the
IdP administrative hierarchy.

#### Implementation Status: implemented

### This System

System-level account managers are documented in the System Security Plan
Appendix A and are drawn from the operations team as designated by the ISSO.

#### Implementation Status: implemented

Parameter Precedence

Parameter values flow through three levels, with each level able to override the previous:

Level	Field	Source	Purpose
1. Catalog	values	Catalog parameter definition	Default/suggested value
2. Profile	profile-values	Profile set-parameters	Baseline-specific tailoring
3. SSP	ssp-values	SSP control markdown header	System-specific implementation value

The SSP-level ssp-values always wins. If ssp-values is set, it overrides both profile-values and catalog values. If ssp-values is empty, the profile value is used. If the profile doesn't set a value, the catalog default applies.

Example showing the same parameter at all three levels:

# Catalog defines: ac-1_prm_1 values: ["organization-defined frequency"]
# Profile sets: profile-values: ["annually"]
# SSP overrides:
x-trestle-set-params:
  ac-1_prm_1:
    values:
      - organization-defined frequency
    profile-values:
      - annually
    ssp-values:
      - at least every 365 days or upon policy change

In the assembled SSP, the resolved value is at least every 365 days or upon policy change.

Handling Multi-Part Controls

OSCAL controls often have statement parts (a, b, c...) and sometimes sub-parts (a.1, a.2). Each maps to a specific markdown structure.

Statement Parts

Each lettered part gets its own ## Implementation for part X. heading:

## Implementation for part a.
### My Component
[Response for part a]

## Implementation for part b.
### My Component
[Response for part b]

Sub-Parts

Nested sub-parts use extended identifiers:

## Implementation for part a.1.
### My Component
[Response for sub-part a.1]

Which Parts Need Responses

• Every statement part must have a response for each component — even if only to say the part is not applicable to that component
• Control statement text, guidance, and objective sections are informational and should not be edited
• Only the sections below the horizontal rule (______________________________________________________________________) accept implementation prose

Compensating Controls

When a control requirement cannot be met as stated, document a compensating control:

1. Set implementation status to alternative:

   #### Implementation Status: alternative
   

2. Describe the compensating control in the implementation prose:

   ### This System
   The system cannot implement centralized session termination as specified
   due to architectural constraints in the legacy middleware layer. As a
   compensating control, the system enforces aggressive idle timeouts
   (15 minutes) at the application layer and requires re-authentication
   for all privileged operations. This compensating control was approved
   by the AO on 2024-06-15 (see POA&M item PM-2024-042).
   

3. Reference the POA&M entry if one exists for tracking the gap

Common Mistakes

Mistake	Impact	Fix
Leaving <!-- Add control implementation --> comments	Assemble silently includes them as empty content — auditors see blank responses	Replace every placeholder with actual prose or remove it
Wrong YAML indentation in headers	Assemble fails with a YAML parse error	Use 2-space indentation consistently, never tabs
Missing ssp-values when values are required	Parameters resolve to catalog defaults which may be vague placeholders	Set explicit ssp-values for every parameter
Writing generic responses like "The system implements this control"	Fails ATO review — assessors need specific detail	Reference specific tools, processes, configurations, and responsible roles
Not addressing every statement part separately	Assemble may silently drop unaddressed parts	Add a response (even "Not applicable to this component") for each part per component
Confusing values vs ssp-values	Assembled SSP uses the wrong parameter value	values = catalog default; ssp-values = what this system actually uses
Editing content above the horizontal rule	Changes are overwritten on the next generate	Only edit content below the ______________________________________________________________________ divider

Where Rules Come From

Rules that appear in SSP markdown (under x-trestle-comp-def-rules) originate from the component definition authoring supply chain — they are not authored in the SSP.

The Rules Supply Chain

CSV spreadsheet (vendor-authored rules, parameters, control mappings)
    --> trestle task csv-to-oscal-cd (produces component-definition JSON)
        --> trestle author component-generate (produces component markdown with rules + editable prose)
            --> Edit markdown: write implementation responses
                --> trestle author component-assemble (merges prose into component-definition JSON)
                    --> trestle author ssp-generate --compdefs ... (pulls rules into SSP markdown)

Rules Are Read-Only in SSP Markdown

The rules section in SSP control markdown is populated from component definitions during ssp-generate. Editing rules directly in SSP markdown has no effect — they will be overwritten on the next ssp-generate run.

To change a rule: Edit the source component definition (CSV or component markdown), reassemble it, then regenerate the SSP.

To Add New Rules

1. Add the rule to the component definition CSV
2. Run trestle task csv-to-oscal-cd to rebuild the component definition
3. Run trestle author component-generate and write the prose response
4. Run trestle author component-assemble
5. Run trestle author ssp-generate to pull the new rule into SSP markdown

For the full two-phase component definition workflow, see trestle-compliance-pipeline.

Cross-References

• trestle-authoring-workflow: The full generate → edit → assemble cycle for SSP control implementation
• trestle-validation: Diagnosing and fixing assemble errors, YAML parse failures, and schema validation issues
• trestle-governance: Enforcing consistent response structure and required headings across control implementations
• trestle-compliance-pipeline: End-to-end pipeline including the rules supply chain and two-phase component definition authoring