Claude Skill

skf-analyze-source

Discover what to skill in a large repo and produce recommended skill briefs. Use when the user requests to "analyze source for skills" or "discover skill opportunities."

LLM Mart · 0 points · 1 views 0 listing impressions 0 install-command copies
Download armelhbobdad-bmad-module-skill-forge-src_skf-analyze-source-484edd0.zip · 85 KB

Install

skills CLI npx skills add https://github.com/armelhbobdad/bmad-module-skill-forge/tree/main/src/skf-analyze-source
Claude Code claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install armelhbobdad-bmad-module-skill-forge@llmmart
Git git clone https://github.com/armelhbobdad/bmad-module-skill-forge.git

The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole armelhbobdad/bmad-module-skill-forge collection as a plugin from our marketplace. Git is the plain clone.

Skill manifest

Analyze Source

Overview

Analyzes a large repo or multi-service project to identify discrete skillable units, map exports and integration points, and produce recommended skill-brief.yaml files as the primary entry point for brownfield onboarding. The analysis must be thorough enough to produce actionable briefs, but scoped enough to avoid overwhelming the user with false positives.

Conventions

  • Bare paths (e.g. references/<name>.md) resolve from the skill root, inside a references/ file too: only a stage's frontmatter nextStepFile, continueFile and nextStepOptions name a file beside it.
  • Module-level path exception: bare paths beginning with knowledge/ or shared/ resolve from the SKF module root ({project-root}/_bmad/skf/ installed, src/ in dev), not the skill root; e.g. shared/health-check.md, which the terminal step chains to.
  • Sibling skills: a path that names another SKF skill's folder (skf-<name>/...) resolves from the SKF module root, and that skill must be installed with this one.
  • {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 source code analyst and decomposition architect collaborating with a developer onboarding an existing project, pairing your codebase-analysis and skill-scoping expertise with their domain knowledge.

Workflow Rules

These rules apply to every step in this workflow:

  • Only load one step file at a time — never preload future steps
  • Run every subagent in the foreground, passing run_in_background: false on every Agent call (Claude Code runs a subagent in the background when the flag is left out), with parallel ones as several such calls in one message, and wait for every long-running command to exit before the next instruction; in a headless run, never end a turn while either still runs, since under claude -p a run stops when its turn ends.
  • Always communicate in {communication_language}; the text persisted into skill-brief.yaml (each recommendation's description and scope.notes) is written in {document_output_language}
  • If {headless_mode} is true, auto-proceed through confirmation gates with their default action, and record each one in the run sink as the gate says

Stages

# Step File Auto-proceed Condition
1 Initialize references/init.md Yes Always
1a Auto-Scope references/step-auto-scope.md Conditional (the coexistence gate) [auto] mode only: bypasses steps 2–6; owns pin resolution, coexistence detection, and the docs-only short-circuit
1b Continue (session resume) references/continue.md Yes An unfinished step-by-step report (mode not 'auto') of the same target and inputs, run again without [auto] (init section 1)
2 Scan Project references/scan-project.md No (confirm) Interactive mode only
3 Identify Units references/identify-units.md No (confirm) Interactive mode only
4 Map & Detect references/map-and-detect.md No (confirm) Interactive mode only
5 Recommend references/recommend.md No (confirm) Interactive mode only
6 Generate Briefs references/generate-briefs.md No (confirm) Interactive mode only
7 Workflow Health Check references/health-check.md Yes Always

Auto mode path: With [auto] present, init routes directly to step 1a, which loads a branch file only when its branch fires: references/step-auto-scope-coexistence.md (a skill already matches), references/step-auto-scope-split.md (a monorepo splits into N briefs) and references/step-auto-scope-corpora.md (a whole-language reference), and routes docs-only targets to references/auto-docs-only.md.

[D] Discover Additional Source (steps 4 and 5) loads references/discover-additional-source.md. Step 2 and [D] make each project path's scan root by references/scan-root.md (steps 3, 4 and 6 make one again when a resumed session no longer finds it), and step 4 and [D] map each unit's exports by references/map-unit-exports.md.

Invocation Contract

Headless callers: the inputs, flags, gates, outputs and exit codes live in references/invocation-contract.md. Interactive runs do not need it.

On Activation

  1. Resolve the envelope helper. {emitEnvelopeHelper} ← the first existing path of {project-root}/_bmad/skf/shared/scripts/skf-emit-result-envelope.py and {project-root}/src/shared/scripts/skf-emit-result-envelope.py. If neither path exists, HARD HALT (exit code 3, halt_reason: "resolution-failure", phase on-activation:helpers) and display only: "skf-emit-result-envelope.py is missing. Nothing was changed. Re-install SKF." In headless mode every other HARD HALT prints its envelope on stderr through the emitter: with the first command below once step 5 has created {run_dir}, before that with the second, the payload on one line. references/headless-contract.md (Halt Envelope) gives the payload rules, and each step file restates them in its opening.

    uv run {emitEnvelopeHelper} emit-halt --workflow skf-analyze-source --run-dir "{run_dir}" --target stderr < "{run_dir}/halt.json"
    uv run {emitEnvelopeHelper} emit-halt --workflow skf-analyze-source --target stderr <<'SKF_ANALYZE_HALT'
    {"phase": "on-activation:config", "reason": "<the halt message>", "halt_reason": "input-missing", "mode": "interactive"}
    SKF_ANALYZE_HALT
    
  2. Load config from {project-root}/_bmad/skf/config.yaml and resolve:

    • project_name, output_folder, user_name, communication_language, document_output_language, forge_data_folder, skills_output_folder, sidecar_path
    • If the config cannot be loaded, HARD HALT (exit code 2, halt_reason: "input-missing", phase on-activation:config): "SKF cannot load {project-root}/_bmad/skf/config.yaml: the workflow has no forge context to run against. Run setup first." When --headless or -H was passed, print its envelope with the second command above.
  3. Resolve {headless_mode}: true if --headless or -H was passed as an argument, or if headless_mode: true in {sidecar_path}/preferences.yaml. Default: false.

  4. Resolve workflow customization. Bind {resolveCustomizationHelper} ← {project-root}/_bmad/skf/shared/scripts/skf-resolve-customization.py, else {project-root}/src/shared/scripts/skf-resolve-customization.py (the first that exists), and run:

    uv run {resolveCustomizationHelper} --skill {skill-root} --project-root {project-root} --key workflow --facts
    

    It merges the bundled {skill-root}/customize.toml with {project-root}/_bmad/custom/skf-analyze-source.toml (team overrides, committed) and .user.toml (personal overrides, gitignored). When it exits non-zero, prints no JSON or is missing, print one line, [activation/warn] customization_resolver_unavailable: <reason> (<reason>: its first stderr line, SKF helper skf-resolve-customization.py not found when the script is missing, no JSON when it printed none). If the resolver cannot run, read {skill-root}/customize.toml alone and use its bundled defaults: the {project-root}/_bmad/custom/ overrides do not apply to this run. In that case keep the reason as {customization_resolver_unavailable} (unset when the resolver ran) for step 5 to record in the run sink.

    Then bind, so stage files use the variable with no conditional at the usage site:

    • {analysisReportTemplatePath} ← workflow.analysis_report_template_path (the bundled templates/analysis-report-template.md when an override leaves it empty)
    • {onCompleteCommand} ← workflow.on_complete (empty: no hook)

    Run workflow.activation_steps_prepend now, and treat workflow.persistent_facts as standing context for the run (the bundled default loads every project-context.md under {project-root}, none in {project-root}/_bmad-output/.skf-run/; if the resolver ran, --facts has applied each ! entry, which drops each earlier entry it names, and expanded each glob, so a file: entry names one file).

  5. Create the run folder. It holds what the run stages for its helpers (the tree listing, each brief's context, the unit records, the halt and result payloads) and the sink of its auto-decisions:

    mkdir -p "{project-root}/_bmad-output/.skf-run" && mktemp -d "{project-root}/_bmad-output/.skf-run/skf-analyze-source-XXXXXXXX"
    

    Bind {run_dir} ← the path it prints. The step that ends the run deletes it once the envelope is out; a HARD HALT keeps it. If it cannot be created, HARD HALT (exit code 4, halt_reason: "write-failed", phase on-activation:run-folder, path {project-root}/_bmad-output/.skf-run): "SKF cannot create its run folder under {project-root}/_bmad-output/.skf-run/: . Nothing was changed." Its envelope goes through the second command of step 1, its payload with "customization_resolver_unavailable": "<reason>" added when step 4 kept one. Once the folder exists and {customization_resolver_unavailable} is set, record the reason in the run sink: write customization_resolver_unavailable: {customization_resolver_unavailable} to {run_dir}/resolver-warning.txt with a file write, never echo (the reason can hold quotes or $( )), then run uv run {emitEnvelopeHelper} record --run-dir "{run_dir}" --warning "$(cat "{run_dir}/resolver-warning.txt")".

  6. Run workflow.activation_steps_append, then load, read the full file, and execute references/init.md to begin the workflow.

Files (bmad-module-skill-forge)
  • assets
    • skill-brief-schema.md 7.4 KB
      Loading…
  • references
    • auto-docs-only.md 8.7 KB
      Loading…
    • continue.md 3.3 KB
      Loading…
    • discover-additional-source.md 12.7 KB
      Loading…
    • generate-briefs.md 18.9 KB
      Loading…
    • headless-contract.md 11.5 KB
      Loading…
    • health-check.md 946 B
      Loading…
    • identify-units.md 16.7 KB
      Loading…
    • init.md 15.1 KB
      Loading…
    • invocation-contract.md 2.5 KB
      Loading…
    • map-and-detect.md 15.9 KB
      Loading…
    • map-unit-exports.md 7.4 KB
      Loading…
    • recommend.md 9.4 KB
      Loading…
    • scan-project.md 8.3 KB
      Loading…
    • scan-root.md 3.3 KB
      Loading…
    • step-auto-scope-coexistence.md 3.6 KB
      Loading…
    • step-auto-scope-corpora.md 4.3 KB
      Loading…
    • step-auto-scope-split.md 8.3 KB
      Loading…
    • step-auto-scope.md 32.4 KB
      Loading…
    • step-shape-detect.md 2.5 KB
      Loading…
    • unit-detection-heuristics.md 12.1 KB
      Loading…
  • templates
    • analysis-report-template.md 595 B
      Loading…
  • customize.toml 3.2 KB
    Loading…
  • SKILL.md 9.7 KB
    ---
    name: skf-analyze-source
    description: Discover what to skill in a large repo and produce recommended skill briefs. Use when the user requests to "analyze source for skills" or "discover skill opportunities."
    ---
    
    # Analyze Source
    
    ## Overview
    
    Analyzes a large repo or multi-service project to identify discrete skillable units, map exports and integration points, and produce recommended skill-brief.yaml files as the primary entry point for brownfield onboarding. The analysis must be thorough enough to produce actionable briefs, but scoped enough to avoid overwhelming the user with false positives.
    
    ## Conventions
    
    - Bare paths (e.g. `references/<name>.md`) resolve from the skill root, inside a `references/` file too: only a stage's frontmatter `nextStepFile`, `continueFile` and `nextStepOptions` name a file beside it.
    - **Module-level path exception:** bare paths beginning with `knowledge/` or `shared/` resolve from the SKF module root (`{project-root}/_bmad/skf/` installed, `src/` in dev), not the skill root; e.g. `shared/health-check.md`, which the terminal step chains to.
    - **Sibling skills:** a path that names another SKF skill's folder (`skf-<name>/...`) resolves from the SKF module root, and that skill must be installed with this one.
    - `{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 source code analyst and decomposition architect collaborating with a developer onboarding an existing project, pairing your codebase-analysis and skill-scoping expertise with their domain knowledge.
    
    ## Workflow Rules
    
    These rules apply to every step in this workflow:
    
    - Only load one step file at a time — never preload future steps
    - Run every subagent in the foreground, passing `run_in_background: false` on every Agent call (Claude Code runs a subagent in the background when the flag is left out), with parallel ones as several such calls in one message, and wait for every long-running command to exit before the next instruction; in a headless run, never end a turn while either still runs, since under `claude -p` a run stops when its turn ends.
    - Always communicate in `{communication_language}`; the text persisted into `skill-brief.yaml` (each recommendation's `description` and `scope.notes`) is written in `{document_output_language}`
    - If `{headless_mode}` is true, auto-proceed through confirmation gates with their default action, and record each one in the run sink as the gate says
    
    ## Stages
    
    | # | Step | File | Auto-proceed | Condition |
    |---|------|------|--------------|-----------|
    | 1 | Initialize | references/init.md | Yes | Always |
    | 1a | Auto-Scope | references/step-auto-scope.md | Conditional (the coexistence gate) | `[auto]` mode only: bypasses steps 2–6; owns pin resolution, coexistence detection, and the docs-only short-circuit |
    | 1b | Continue (session resume) | references/continue.md | Yes | An unfinished step-by-step report (`mode` not `'auto'`) of the same target and inputs, run again without `[auto]` (init section 1) |
    | 2 | Scan Project | references/scan-project.md | No (confirm) | Interactive mode only |
    | 3 | Identify Units | references/identify-units.md | No (confirm) | Interactive mode only |
    | 4 | Map & Detect | references/map-and-detect.md | No (confirm) | Interactive mode only |
    | 5 | Recommend | references/recommend.md | No (confirm) | Interactive mode only |
    | 6 | Generate Briefs | references/generate-briefs.md | No (confirm) | Interactive mode only |
    | 7 | Workflow Health Check | references/health-check.md | Yes | Always |
    
    **Auto mode path:** With `[auto]` present, init routes directly to step 1a, which loads a branch file only when its branch fires: `references/step-auto-scope-coexistence.md` (a skill already matches), `references/step-auto-scope-split.md` (a monorepo splits into N briefs) and `references/step-auto-scope-corpora.md` (a whole-language reference), and routes docs-only targets to `references/auto-docs-only.md`.
    
    **[D] Discover Additional Source** (steps 4 and 5) loads `references/discover-additional-source.md`. Step 2 and [D] make each project path's scan root by `references/scan-root.md` (steps 3, 4 and 6 make one again when a resumed session no longer finds it), and step 4 and [D] map each unit's exports by `references/map-unit-exports.md`.
    
    ## Invocation Contract
    
    Headless callers: the inputs, flags, gates, outputs and exit codes live in `references/invocation-contract.md`. Interactive runs do not need it.
    
    ## On Activation
    
    1. **Resolve the envelope helper.** `{emitEnvelopeHelper}` ← the first existing path of `{project-root}/_bmad/skf/shared/scripts/skf-emit-result-envelope.py` and `{project-root}/src/shared/scripts/skf-emit-result-envelope.py`. If neither path exists, HARD HALT (exit code 3, `halt_reason: "resolution-failure"`, phase `on-activation:helpers`) and display only: "`skf-emit-result-envelope.py` is missing. Nothing was changed. Re-install SKF." In headless mode every other HARD HALT prints its envelope on stderr through the emitter: with the first command below once step 5 has created `{run_dir}`, before that with the second, the payload on one line. `references/headless-contract.md` (Halt Envelope) gives the payload rules, and each step file restates them in its opening.
    
       ```bash
       uv run {emitEnvelopeHelper} emit-halt --workflow skf-analyze-source --run-dir "{run_dir}" --target stderr < "{run_dir}/halt.json"
       uv run {emitEnvelopeHelper} emit-halt --workflow skf-analyze-source --target stderr <<'SKF_ANALYZE_HALT'
       {"phase": "on-activation:config", "reason": "<the halt message>", "halt_reason": "input-missing", "mode": "interactive"}
       SKF_ANALYZE_HALT
       ```
    
    2. Load config from `{project-root}/_bmad/skf/config.yaml` and resolve:
       - `project_name`, `output_folder`, `user_name`, `communication_language`, `document_output_language`, `forge_data_folder`, `skills_output_folder`, `sidecar_path`
       - If the config cannot be loaded, HARD HALT (exit code 2, `halt_reason: "input-missing"`, phase `on-activation:config`): "SKF cannot load `{project-root}/_bmad/skf/config.yaml`: the workflow has no forge context to run against. Run setup first." When `--headless` or `-H` was passed, print its envelope with the second command above.
    
    3. **Resolve `{headless_mode}`**: true if `--headless` or `-H` was passed as an argument, or if `headless_mode: true` in `{sidecar_path}/preferences.yaml`. Default: false.
    
    4. **Resolve workflow customization.** Bind `{resolveCustomizationHelper}` ← `{project-root}/_bmad/skf/shared/scripts/skf-resolve-customization.py`, else `{project-root}/src/shared/scripts/skf-resolve-customization.py` (the first that exists), and run:
    
       ```bash
       uv run {resolveCustomizationHelper} --skill {skill-root} --project-root {project-root} --key workflow --facts
       ```
    
       It merges the bundled `{skill-root}/customize.toml` with `{project-root}/_bmad/custom/skf-analyze-source.toml` (team overrides, committed) and `.user.toml` (personal overrides, gitignored). When it exits non-zero, prints no JSON or is missing, print one line, `[activation/warn] customization_resolver_unavailable: <reason>` (`<reason>`: its first stderr line, `SKF helper skf-resolve-customization.py not found` when the script is missing, `no JSON` when it printed none). If the resolver cannot run, read `{skill-root}/customize.toml` alone and use its bundled defaults: the `{project-root}/_bmad/custom/` overrides do not apply to this run. In that case keep the reason as `{customization_resolver_unavailable}` (unset when the resolver ran) for step 5 to record in the run sink.
    
       Then bind, so stage files use the variable with no conditional at the usage site:
    
       - `{analysisReportTemplatePath}` ← `workflow.analysis_report_template_path` (the bundled `templates/analysis-report-template.md` when an override leaves it empty)
       - `{onCompleteCommand}` ← `workflow.on_complete` (empty: no hook)
    
       Run `workflow.activation_steps_prepend` now, and treat `workflow.persistent_facts` as standing context for the run (the bundled default loads every `project-context.md` under `{project-root}`, none in `{project-root}/_bmad-output/.skf-run/`; if the resolver ran, `--facts` has applied each `!` entry, which drops each earlier entry it names, and expanded each glob, so a `file:` entry names one file).
    
    5. **Create the run folder.** It holds what the run stages for its helpers (the tree listing, each brief's context, the unit records, the halt and result payloads) and the sink of its auto-decisions:
    
       ```bash
       mkdir -p "{project-root}/_bmad-output/.skf-run" && mktemp -d "{project-root}/_bmad-output/.skf-run/skf-analyze-source-XXXXXXXX"
       ```
    
       Bind `{run_dir}` ← the path it prints. The step that ends the run deletes it once the envelope is out; a HARD HALT keeps it. If it cannot be created, HARD HALT (exit code 4, `halt_reason: "write-failed"`, phase `on-activation:run-folder`, path `{project-root}/_bmad-output/.skf-run`): "SKF cannot create its run folder under `{project-root}/_bmad-output/.skf-run/`: {the first stderr line}. Nothing was changed." Its envelope goes through the second command of step 1, its payload with `"customization_resolver_unavailable": "<reason>"` added when step 4 kept one. Once the folder exists and `{customization_resolver_unavailable}` is set, record the reason in the run sink: write `customization_resolver_unavailable: {customization_resolver_unavailable}` to `{run_dir}/resolver-warning.txt` with a file write, never `echo` (the reason can hold quotes or `$( )`), then run `uv run {emitEnvelopeHelper} record --run-dir "{run_dir}" --warning "$(cat "{run_dir}/resolver-warning.txt")"`.
    
    6. Run `workflow.activation_steps_append`, then load, read the full file, and execute `references/init.md` to begin the workflow.
    

Comments (0)

Sign in to join the conversation.

No comments yet.

Reviews (0)

No reviews yet.