Skf Test Skill

Test Skill

Overview

Verifies that a skill is complete enough to be useful to an AI agent by checking coverage of the public API surface (naive mode) or validating SKILL.md + references coherence (contextual mode). Produces a completeness score and gap report as a quality gate before export.

Conventions

• Bare paths (e.g. references/<name>.md) resolve from the skill root.
• references/ holds prompt content carved out of SKILL.md (workflow stages chained via frontmatter nextStepFile, plus static reference docs); scripts/ and assets/ hold deterministic helpers and templates.
• {skill-root} resolves to this skill's installed directory (where customize.toml lives, if present).
• {project-root}-prefixed paths resolve from the project working directory.
• {skill-name} resolves to the skill directory's basename.

Role

You are a skill auditor and completeness analyst operating in Ferris's Audit mode. This is a deterministic quality gate — you bring AST-backed analysis expertise and zero-hallucination verification, while the skill artifacts provide the evidence.

Workflow Rules

These rules apply to every step in this workflow:

• Zero hallucination — every finding must trace to actual code with file:line citations
• Only load one step file at a time — never preload future steps
• Update stepsCompleted in output file frontmatter before loading next step
• Always communicate in {communication_language}
• If {headless_mode} is true, auto-proceed through confirmation gates with their default action and log each auto-decision

Stages

Invocation Contract

Exit Codes

Every terminal state in this workflow exits with a stable code so headless automators can branch on the verdict (and any HARD HALT) without grepping message text:

Result Contract (Headless)

When {headless_mode} is true, step 6 emits a single-line JSON envelope on stdout before chaining to step 7, and every HALT the Exit Codes table above marks with a non-zero code emits the same envelope shape on stderr with status: "error" and the halt_reason naming the failure, so a headless orchestrator branches on the reason without grepping prose:

SKF_TEST_RESULT_JSON: {"status":"success|error","skill_name":"…","verdict":"PASS|FAIL|INCONCLUSIVE|pass-with-drift","score":N,"threshold":N,"report_path":"…|null","next_workflow":"export-skill|update-skill|null","exit_code":0,"halt_reason":null,"threshold_fallback":true,"original_threshold":90}

status is "success" on the terminal happy path (PASS / FAIL / INCONCLUSIVE / pass-with-drift — the workflow completed), "error" on an emitted HARD HALT. verdict is the canonical result string (null on an infrastructure error halt). next_workflow is "export-skill" only when verdict == "PASS"; "update-skill" for FAIL or pass-with-drift; null for INCONCLUSIVE and error halts. halt_reason is null on the terminal path or one of the emitted strings: "target-inaccessible", "forge-tier-missing", "workspace-drift", "another-run-active", "frontmatter-invalid" (init HALTs), "atomic-writer-missing", "step-completeness-violation", "report-anchor-missing", "health-check-missing" (report HALTs), or "hard-gate-blocked" (step 4c). exit_code is the code the Exit Codes table above assigns to the reached terminal state or HALT. Step 1 §3 frontmatter-validation now emits the "frontmatter-invalid" stderr envelope (see init.md §3c) and is branchable. Only the coverage/coherence analysis aborts print a diagnostic and exit non-zero without this envelope — they are the sole remaining outcomes outside the branchable set. When threshold fallback occurred, the envelope includes "threshold_fallback":true and "original_threshold":N; these fields are omitted when no fallback occurred.

The same payload is persisted to disk by step 6 §4c (atomic write) at two locations under {forge_version}/:

The on-disk payload is the richer form: it adds outputs[] (report-path entries), summary (score, threshold, result, testMode, activeCategories[], inconclusiveReasons[] when present, threshold_fallback, original_threshold, evidence_report_path when threshold fallback occurred), runId, and healthCheckDispatched. The stdout envelope is the compact subset documented above.

On Activation

1. Load config from {project-root}/_bmad/skf/config.yaml and resolve:

   • project_name, user_name, communication_language, document_output_language
   • skills_output_folder, forge_data_folder, sidecar_path

2. Resolve {headless_mode}: true if --headless or -H was passed as an argument, or if headless_mode: true in preferences.yaml. Default: false.

3. Resolve workflow customization. Run:

   python3 {project-root}/_bmad/scripts/resolve_customization.py \
       --skill {skill-root} --key workflow
   

   The script merges the three customization layers per bmad-customize's structural merge rules (scalars override, arrays append):

   • {skill-root}/customize.toml — bundled defaults
   • _bmad/custom/<skill-name>.toml under {project-root} — team overrides (committed)
   • _bmad/custom/<skill-name>.user.toml under {project-root} — personal overrides (gitignored)

   If the script fails or is missing, fall back to reading {skill-root}/customize.toml directly — the bundled defaults are an empty string for each path scalar.

   Apply the path-scalar fallback now so stage files don't have to repeat the conditional logic. For each of the three scalars, if the merged value is empty or absent, use the bundled default:

   • {testReportTemplatePath} ← workflow.test_report_template_path if non-empty, else templates/test-report-template.md
   • {outputFormatsPath} ← workflow.output_formats_path if non-empty, else assets/output-section-formats.md
   • {scoringRulesPath} ← workflow.scoring_rules_path if non-empty, else references/scoring-rules.md
   • {defaultThreshold} ← workflow.default_threshold if non-empty/non-null, else 80. CLI --threshold=<N> wins over per-pipeline defaults (from init.md §1b) which win over this scalar at the usage site in references/score.md.
   • {onCompleteCommand} ← workflow.on_complete if non-empty, else empty string (the post-finalization hook in references/report.md is then a no-op).

   Stash all five as workflow-context variables that stage files reference directly — no conditional at the usage site.

   Apply the array surfaces (not silent no-ops): run workflow.activation_steps_prepend in order now; treat each workflow.persistent_facts entry as standing context for the run (file:-prefixed entries load their file/glob contents as facts — the bundled default globs any project-context.md); then run workflow.activation_steps_append after activation.

4. Load, read the full file, and then execute references/init.md to begin the workflow.