s0912758806p/agentic-sop-to-work

ALCOA Guard checks records or GMP documents for ALCOA+ data integrity issues such as attribution, backdating, completeness, accuracy, and consistency. It returns a draft for human review and only verifies what can be checked mechanically.

Agentic SOP helps turn a human SOP or manual process into an agentic workflow, including new agent tools and workflow projects. It is used when you need to draft, decompose, and wire a process into guarded steps before implementation.

Agentic Workflow Audit reviews an existing agent or LLM pipeline to see whether it is truly decomposed into bounded tasks with explicit contracts, checks, and recovery paths. Use it when auditing workflow structure, task boundaries, or mega-agent drift.

Bad Plugin provides an agentic workflow plugin for Claude Code, oriented around SOP-driven tasks. Use it when you want to adapt or automate a structured agent workflow in Python.

Compute reads a readings artifact and calculates count, sum, mean, min, and max while preserving the upstream trace. Use it when you need a stats artifact for a downstream step.

Extract parses plain-text lines like key: number into a structured readings artifact with source-location trace. Use it when you need a simple, reusable first-step extractor for downstream processing.

Good Plugin is a Claude Code plugin for agentic workflow automation. Use it when you want to run SOP-driven tasks with Claude Code.

Plugin Forge scaffolds Claude Code plugins and lints a single plugin or the whole marketplace for manifest, frontmatter, stdlib-only, and test-harness rules. Use it when creating a new plugin or checking existing plugins against house invariants.

Report renders a draft Markdown summary from a stats artifact and includes a source-trace section. Use it when you need a human-reviewed report from upstream stats.

Second Opinion checks a finished draft’s claims against its cited evidence and flags fabrication, overreach, or invented IDs. Use it to sanity-check a report before sign-off or to review a document when you need an independent second read.

Six-Rung Ladder filters decisions about whether to write code, add a dependency, or introduce an abstraction. It walks through YAGNI, stdlib, platform features, existing dependencies, one-line fixes, and the smallest working version.