Cybersecurity knowledge map
Controls with a project attached
Conceptual mappings from AI/engineering patterns to security language. They are not claimed as perfect one-to-one industry equivalences.
AI / engineering
Agent permissions
Cybersecurity
Least privilege
The triage agent receives only the tools required for read-only collection. Remediation authority is a separate, human-held capability.
Used in: Home-Lab Endpoint Triage Agent · Knowledge Publishing Agent
AI / engineering
Tool allowlist
Cybersecurity
Application control / authorized capability boundary
Approved tools are enumerated. Completing a task more easily with an out-of-policy tool is treated as a failed control, not a feature.
Used in: Home-Lab Endpoint Triage Agent · Knowledge Publishing Agent · Clearward
AI / engineering
Audit trail of tool calls, retries, and approvals
Cybersecurity
Security logging and accountability
If a manipulation attempt is not visible afterward, the control did not hold.
Used in: Home-Lab Endpoint Triage Agent · Knowledge Publishing Agent · Basaltborne
AI / engineering
Output validation in code
Cybersecurity
Verify before trust
Required fields, hashes, schema, and style checks are deterministic. Model judgment is not allowed to certify its own correctness.
Used in: Knowledge Publishing Agent · Clearward · Basaltborne · TinyClaw
AI / engineering
Isolated lab environment
Cybersecurity
Security segmentation
Adversarial testing of the triage agent started only after the VM had no internet, no LAN path, and no host passthrough.
Used in: Home-Lab Endpoint Triage Agent
AI / engineering
Secrets and credentials stay human-only
Cybersecurity
Credential protection
Clearward agent instructions forbid collecting passwords, recovery codes, payment-card data, or MFA secrets. Basaltborne keeps service-role credentials server-only.
Used in: Clearward · Basaltborne
AI / engineering
Automated tests as merge gates
Cybersecurity
Preventive / detective controls
Basaltborne runs lint, typecheck, Vitest, pgTAP, Playwright, and migration replay as independent GitHub Actions checks.
Used in: Basaltborne · Clearward
AI / engineering
Rollback planned before change
Cybersecurity
Recovery / incident response
Recommended triage actions ship with expected results, validation criteria, operational risk, and a rollback procedure. Hyper-V checkpoints revert test cycles.
Used in: Home-Lab Endpoint Triage Agent · Basaltborne
AI / engineering
CI pipeline on protected main
Cybersecurity
Automated change validation
Nothing merges without the suite. Branch protection is the technical backstop for the same human-approval principle used in the products.
Used in: Basaltborne
AI / engineering
Trusted retrieval and source verification
Cybersecurity
Trusted data boundary / provenance control
TinyClaw normalizes and timestamps sources before the model sees them. The Jeep assistant is instructed not to present forum consensus as verified specification.
Used in: TinyClaw · Domain-Specific AI Expert · Basaltborne
AI / engineering
Human approval gate
Cybersecurity
Privileged change authorization
Publication, isolation, process termination, account disablement, and customer release are privileged transitions. The model recommends; it does not act.
Used in: Home-Lab Endpoint Triage Agent · Knowledge Publishing Agent · Clearward · Basaltborne · Multi-Stage Knowledge Publishing Workflow
AI / engineering
Untrusted logs, documents, and command output
Cybersecurity
Untrusted-input handling
Content may contain instructions intended to manipulate the model. Tool policy overrides anything asserted in the content.
Used in: Home-Lab Endpoint Triage Agent · Knowledge Publishing Agent
Security concepts × evidence
IAM / authorization
Least privilege
Secure SDLC
Logging and accountability
Validation
Change control
Segmentation
Secrets management
Endpoint security
Incident response / recovery
Threat modeling the agent
Application security
Security automation
Physical / operational security systems
IAM / authorization
Who can do what is assigned outside the model and enforced at the boundary that matters.
Where learned: Basaltborne RLS and RPCs; Clearward human-only credential boundary; triage least-privilege account.
Evidence: Service-role credentials never shipped to the browser. Agent account is collection-only.
Applied in: Basaltborne · Clearward · Home-Lab Endpoint Triage Agent
Least privilege
Separate collection, analysis, and remediation. Give the current task only what it needs.
Where learned: Triage agent; publishing-agent allowlist; Clearward directory and spending boundaries.
Evidence: Consequential actions always require explicit analyst or operator approval.
Applied in: Home-Lab Endpoint Triage Agent · Knowledge Publishing Agent · Clearward
Secure SDLC
Phase gates, PR review, protected main, and tests that run somewhere other than a laptop.
Where learned: Basaltborne feature branches, acceptance-criteria commits, CI-protected main.
Evidence: AI-authored changes still require an explicit review pass against tests and criteria.
Applied in: Basaltborne
Logging and accountability
Decisions, tool calls, retries, and approvals have to be reconstructable.
Where learned: Triage adversarial study scored whether attempts appeared in the audit log.
Evidence: Publishing-agent design requires an audit trail of disposition plus versioned artifacts.
Applied in: Home-Lab Endpoint Triage Agent · Knowledge Publishing Agent · Basaltborne
Validation
Schema, hashes, required fields, RLS tests, and visual review — not the generator marking its own homework.
Where learned: Clearward formula/package QA; Basaltborne Zod + pgTAP; TinyClaw source checks.
Evidence: 49 automated checks on Pet Care System v1.0; 120/120 pgTAP on Basaltborne.
Applied in: Clearward · Basaltborne · TinyClaw · Knowledge Publishing Agent
Change control
Publication, packaging, and deploy are privileged transitions with a human on the gate.
Where learned: Clearward release architecture; Basaltborne draft vs published; Self Help human review.
Evidence: Nothing reached ServiceNow unreviewed; customer packages require approval and hashes.
Applied in: Clearward · Basaltborne · Multi-Stage Knowledge Publishing Workflow
Segmentation
Assume the test can succeed, and make sure success stays inside the box.
Where learned: Triage VM: private virtual switch, no internet, no LAN, no host passthrough.
Evidence: No adversarial testing started until the environment was contained.
Applied in: Home-Lab Endpoint Triage Agent
Secrets management
Credentials are a human-only boundary. Service roles stay server-side. Prompts are not a vault.
Where learned: Clearward AGENTS.md; Basaltborne server-only privilege.
Evidence: Never ask for passwords, recovery codes, payment-card data, or MFA secrets.
Applied in: Clearward · Basaltborne
Endpoint security
Read-only collection of processes, persistence, network, logs, hashes — then a human.
Where learned: Home-lab triage agent on an isolated Windows VM.
Evidence: Evidence scope is deliberately read-only; remediation is out of band.
Applied in: Home-Lab Endpoint Triage Agent
Incident response / recovery
Recommended actions include expected results, validation, operational risk, and rollback. Checkpoints revert test cycles.
Where learned: Triage recommendations and Hyper-V checkpoints; Basaltborne planned undo for migrations.
Evidence: Checkpoints taken before every test cycle so state could be rolled back to known-good.
Applied in: Home-Lab Endpoint Triage Agent · Basaltborne
Threat modeling the agent
The agent is in the threat model: injected logs, deceptive names, extra-tool prompts, contradictory evidence, broken tools.
Where learned: Documented adversarial study against the triage agent.
Evidence: Tactics exercised and recorded with observed behavior, control result, and fix.
Applied in: Home-Lab Endpoint Triage Agent
Application security
Untrusted browser, validated RPCs, private tables, fail closed, information minimization.
Where learned: Basaltborne database-security and public-API design.
Evidence: Internal publisher identity does not leak through public content APIs.
Applied in: Basaltborne
Security automation
Accelerate gathering and documentation without substituting the model for judgment.
Where learned: Stated cybersecurity application of this portfolio: analyst control remains the point.
Evidence: Roadmap phases 3–5 (collector, reasoning, adversarial testing) marked complete for the lab build.
Applied in: Home-Lab Endpoint Triage Agent · TinyClaw
Physical / operational security systems
Operational security thinking from service-desk work and from treating a lab VM as hostile rather than trusted.
Where learned: Professional context at Chelan County PUD; host-side separation during triage testing.
Evidence: Distinct admin account and host firewall rules, treating the VM as hostile.
Applied in: Multi-Stage Knowledge Publishing Workflow · Home-Lab Endpoint Triage Agent