readyos-vice-dotnet-artifact-analyst - SKILL.md Agent Skill

name: readyos-vice-dotnet-artifact-analyst description: Run or reuse ReadyOS .NET VICE task runs, then analyze generated vice_auto artifacts (manifest, stage outputs, screen/state/memory/REU dumps, register captures) and produce crash-path summaries that complement readyos-stability-analyst.

ReadyOS VICE Dotnet Artifact Analyst

Use this skill when the user wants analysis centered on .NET ViceTasks automation artifacts (logs/vice_auto_*) and wants deterministic crash-path evidence from generated run outputs.

Trigger Signals

"use the dotnet VICE tool"
"analyze latest vice_auto artifacts"
"compare two vice_auto runs"
"manifest/degraded steps/register dump/debug ring"
"post_input drops to BASIC"

Hard Rules

Prefer artifact-first analysis from logs/vice_auto_*.
Use .NET ViceTasks (tools/vice_tasks_dotnet) for automated capture when a fresh run is required.
Treat manifest.json + stage artifacts as primary truth; treat generated summaries as secondary.
Always label ReadyShell overlay profile in conclusions:
release/default (READYSHELL_PARSE_TRACE_DEBUG=0): READYSHELL_OVERLAYSIZE=0x3800, __OVERLAYSTART__=0x8E00
debug trace (READYSHELL_PARSE_TRACE_DEBUG=1): READYSHELL_OVERLAYSIZE=0x3B00, __OVERLAYSTART__=0x8B00
Treat ReadyShell REU ownership as dynamic. Resolve overlay/state/debug banks from logical REU bank 0, generated ReadyShell metadata, or current $C600-$C6FF type-table dumps; do not assume fixed physical 0x40/0x41/0x43.
When stability-level interpretation is requested, run $readyos-stability-analyst against the same run root.

Workflow

If a run root is provided, analyze it directly.
Otherwise, run the probe plan with .NET ViceTasks and analyze the latest run.
Produce a concise artifact summary:

overall status, failed/degraded steps
key stage outcomes (verify_1_echo, check_prompt_post_input, post-input captures)
post-input screen state (RS> vs READY.)
register capture (debug_post_input_regs)
RAM debug ring/head ($C7A0/$C7F0) and profile-aware overlay window samples ($8E00 release / $8B00 debug)

Optionally compare against a baseline run.
If needed, chain to $readyos-stability-analyst for contract-level findings.

Commands

Run (or reuse latest) and summarize: skills/readyos-vice-dotnet-artifact-analyst/scripts/run_dotnet_probe_and_analyze.sh
Run with explicit ReadyShell parse-trace profile: skills/readyos-vice-dotnet-artifact-analyst/scripts/run_dotnet_probe_and_analyze.sh --parse-trace-debug 0 skills/readyos-vice-dotnet-artifact-analyst/scripts/run_dotnet_probe_and_analyze.sh --parse-trace-debug 1
Analyze a specific run root: python3 skills/readyos-vice-dotnet-artifact-analyst/scripts/summarize_dotnet_vice_artifacts.py --run-root logs/vice_auto_YYYYMMDD_HHMMSS
Compare two runs: python3 skills/readyos-vice-dotnet-artifact-analyst/scripts/summarize_dotnet_vice_artifacts.py --run-root <new> --compare-root <old>

References

Artifact interpretation checklist: skills/readyos-vice-dotnet-artifact-analyst/references/artifact_playbook.md
C64/cc65 web research toolkit: skills/readyos-vice-dotnet-artifact-analyst/references/c64_cc65_web_sources.md
.NET ViceTasks behavior: tools/vice_tasks_dotnet/README.md
Stability companion skill: skills/readyos-stability-analyst/SKILL.md