Slash Command

/myco-autolyze

Sweeps codebase for stale narrative references (version drift, paths, identifiers, numeric/test drifts) using autolysis subagent. Returns patch table; does not apply.

Python

Pytest

code-quality

npx claudepluginhub battam1111/myco --plugin myco

Details

Argument[category-filter]

Prompt Preview

The user wants a stale-narrative sweep. Argument (if any): $ARGUMENTS

If $ARGUMENTS is empty: full sweep across all 5 categories (stale_version / stale_path / stale_identifier / numeric_drift / test_count).

If $ARGUMENTS is one of the category names: scope the sweep to that category only.

Dispatch the **autolysis** subagent (defined at `.claude/agents/autolysis.md`). Pass the category filter (if any) as the dynamic input. The subagent will:

1. Establish ground truth: read live `__version__`, verb count from manifest, dim count from `_canon_lint.yaml`, test count from `pytest --collect-o...

Command Content

Other plugins with /myco-autolyze

/docs-technical-review

Validates documentation technical accuracy against code repositories, detecting stale commands, flags, APIs, config keys, code examples, and file paths. Produces report (default) or auto-fixes issues with --fix.

9 tools

docs-tools

/clean

119

Analyzes recent changes in commit range, unstaged files, or context for dead code, orphaned artifacts from failed branches; generates validated cleanup tasks and reports.

spectre

/hatch3r-revision

Revises agent-implemented code in fresh context window: reconstructs from git diff, interviews user for feedback, fixes issues, cleans leftovers, runs quality pipeline toward merge readiness.

hatch3r

/legacy

Characterizes legacy codebases, adds characterization/golden master tests, audits dependencies, assesses obsolescence, detects dead code, generates roadmaps in docs/legacy/. Supports flags like --assess, --characterize <path>.

godmode

Stats

Stars58

Forks8

Last CommitApr 28, 2026

Actions

View Source View Plugin View on GitHub View README

Help us improve

Share bugs, ideas, or general feedback.

The user wants a stale-narrative sweep. Argument (if any): $ARGUMENTS If $ARGUMENTS is empty: full sweep across all 5 categories (stale_version / stale_path / stale_identifier / numeric_drift / test_count). If $ARGUMENTS is one of the category names: scope the sweep to that category only. Dispatch the **autolysis** subagent (defined at `.claude/agents/autolysis.md`). Pass the category filter (if any) as the dynamic input. The subagent will: 1. Establish ground truth: read live `__version__`, verb count from manifest, dim count from `_canon_lint.yaml`, test count from `pytest --collect-o...

The user wants a stale-narrative sweep. Argument (if any): $ARGUMENTS

If $ARGUMENTS is empty: full sweep across all 5 categories (stale_version / stale_path / stale_identifier / numeric_drift / test_count).

If $ARGUMENTS is one of the category names: scope the sweep to that category only.

Dispatch the autolysis subagent (defined at .claude/agents/autolysis.md). Pass the category filter (if any) as the dynamic input. The subagent will:

Establish ground truth: read live __version__, verb count from manifest, dim count from _canon_lint.yaml, test count from pytest --collect-only, subsystem count from _canon.yaml::subsystems.
Run targeted Greps within the category scope, honoring the exclusion list (legacy_v0_3, notes/integrated/n_*, contract_changelog past sections, _excreted, etc.).
For each candidate, Read for context to confirm true staleness.
Produce a patch table: file:line:stale_string → replacement, classified by category.
Cap output at 100 findings; if cap hit, recommend the user split by category.
Surface (without proposing patches) any L0/L1/L2 doctrine hits — those require a craft, not a sweep.

When autolysis returns, relay the patch table verbatim to the user.

Do NOT apply any of the proposed patches yourself. The user reviews and approves (or applies in a separate turn). The patch table is the deliverable.