PFC5 Core Modeling: Securing Asphalt Simulation Foundations

Where this fits

You’re an agent that must rebuild or audit an asphalt specimen model in PFC 5.0. You have legacy scripts, mixed version syntax, and unclear save states. Without a disciplined foundation, your simulation drifts into unstable geometry and unverifiable results. This skill enforces a version-pure, staged workflow for building and auditing the core model structure—domains, walls, balls, clumps, groups, ranges, and contact assignments—so your agent can start or inspect a case confidently.

Why agents benefit

• Detects unsupported-version syntax in PFC5 source trees, preventing silent command misparses.
• Classifies model files into accepted, restart-only, diagnostic-only, failed, and obsolete states, letting agents decide next actions reliably.
• Structures build/compaction/contact/save stages explicitly, enabling reproducible checkpoints and rollback.
• Enforces top-level cmat and group/range organization, so agent-driven parameter changes don’t corrupt contact logic.
• Provides explicit handoff routing to specialized skills (e.g., gradation, FISH, servo) without mixing responsibilities.

Practical scenarios

Auditing a PFC5 case from a colleague
An agent receives a directory of PFC5 files. It runs the skill’s audit mode, identifies mixed 6.0 syntax (like wall generate box instead of wall create), and flags files that are only restart-compatible. It then produces a prioritized list of fixes before any simulation runs.

Building a fresh asphalt specimen for a mix design
The agent must create a 2D or 3D domain, populate it with clumps representing aggregate, define groups for binder and air voids, and assign a linear contact model at the top level. The skill guides the agent through the required stage order, saves each stage with a clear naming convention, and validates that no future-only syntax leaks in.

Recovering a mid-project case after a failed compaction run
After a server crash, the agent needs to determine which saved stage is valid to resume from. The skill’s state classifier checks file hashes and stage metadata, suggests whether to restart from the last saved equilibrium or rerun compensation, and prevents accidental use of a diagnostic-only save.

Add it to your agent workflow

Install via your skill manager:

npx skills add Echo-aloha/asphalt-codex-skills-5 --skill pfc5-core-modeling

Then, in your agent’s plan, invoke the skill with required inputs, e.g.:

pfc5-core-modeling --mode=audit --source=./case_dir --target-version=5.0

The skill returns a structured report of version violations, stage completeness, and recommended next actions. For more context, see the skill’s entry on OpenAgentSkill.

Compare before adopting

Before committing, weigh quality signals: does the skill’s README include concrete expected outputs and edge cases? Check maintenance freshness—last commit and open issues. Compare against alternatives like a generic PFC automation library that lacks version-purity enforcement. Also consider how well your agent’s existing workflow handles staged saves and file classification; if you already have custom scripts, this skill must integrate cleanly.

Why it is worth tracking

This skill addresses a real pain point: PFC version drift in asphalt modeling scripts. With 31 stars and an active author, it shows early community traction. It fills a niche between raw command generation and case-handoff, making it a valuable checkpoint for any agent that must produce reproducible, audit-ready PFC5 models. Track it if you frequently handle legacy PFC cases or need to enforce version-pure builds in agent-driven simulation pipelines.