All skills

Assess documentation for gaps, staleness, quality, and files outside the defined doc set. Preserve useful knowledge in canonical docs, then remove superseded or unnecessary files. Check changed code when the tree is dirty, otherwise assess the whole repo. Explicit modes: --review, --update, --generate, or --session to capture durable knowledge from a conversation or transcript. Use for doc creation, cleanup, freshness, quality, or saving occasional manual test procedures and results whose numbers a later run will compare against; routine suite runs and ordinary QA passes do not need test records.

Use this Skill: https://skilld.dev/gh/nielsmadan/agentic-coding/doc

This session only. Nothing lands on disk.

referencesgenerate-templates.md

≈1k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Documentation Templates

Each doc separates how it works now (current-state — --update rewrites this as code changes) from a short why (decisions/rationale — rarely touched). Don't restate code: reference signatures by file:line and capture what the agent can't derive by reading the source.

Module/Service Template

# {Module Name}

## Purpose
{What this module does and why it exists — 1-3 sentences}

## How it works (current state)
{How it behaves now: the flow, key responsibilities, how callers use it.
Reference the code as canonical — do NOT paste signatures.}

## Key entry points
- `lib/services/foo.dart:42` — {what this is, its contract / when to call it}
- `lib/services/foo.dart:88` — {…}

## Gotchas
{Non-obvious behavior, common mistakes, platform quirks}

## Why
{Only non-obvious decisions: why this approach over the alternative. Omit if none.}

## Related
{Links to related docs}

Feature Template

# {Feature Name}

## Overview
{What the feature does for users — 1-3 sentences}

## How it works (current state)
{The implementation as it stands now: key components and how they fit together.}

## Key entry points
- `lib/features/x/…:NN` — {role of this file}

## Configuration
{Config options that affect behavior, if any}

## Gotchas
{Edge cases, limitations, common issues}

## Why
{Only non-obvious decisions/tradeoffs. Omit if none.}

ADR Template (docs/decisions/NNNN-<slug>.md)

One file per decision, numbered in order. Records what was chosen and why at the time — an append-only record, never rewritten to match today's code (see "Doc lifecycles" in principles.md).

# {NNNN} — {the decision, stated as a claim}

**Status:** accepted ({YYYY-MM-DD})

## Context
{The forces that made this a decision rather than an obvious call: the constraint, the thing
that broke, the two options that both looked reasonable. Link the ADRs it builds on.}

## Decision
{What we chose, present tense. One paragraph.}

## Consequences
{What it buys and what it costs — including the bad parts: what is now harder, what a future
reader will trip over, what this forecloses.}

When it is later replaced, the status becomes superseded by [0012](0012-<slug>.md) and the new ADR opens with Supersedes [0004](0004-<slug>.md). A partial replacement keeps the status accepted and adds a Superseded in part by [0007](0007-<slug>.md), which {what moved}. bullet under Consequences. Nothing else in the file changes.

Reference Template (docs/reference/<subject>.md)

For an external subject — a dependency, platform, harness, protocol, or service. See "External reference" in principles.md for when one is warranted.

# {Subject}

{One line: what it is and why we build against it.}

## What we verified
{Claims we established ourselves. Each carries the date AND the version probed, plus a
one-line note on how — so the next reader can re-run it.}

- **{Behavior}** — {what happens}. Verified {YYYY-MM-DD} against {name} {version}.
  {The probe, in one line.}

## What upstream documents
{Claims taken from the docs, each with its link. Kept separate from the above: these are
the ones upstream can change without telling us.}

## Gotchas
{Surprises, contradictions with the docs, footguns, things that look supported and aren't.}

## How this affects us
{Only the consequence for this repo — the workaround we adopted, the constraint it puts on
our design. Point at the code (`file:line`); don't restate it.}

## Upstream
- Docs: {url}
- Changelog / releases: {url}
- Relevant issues: {url}

Notes:

  • For files longer than ~100 lines, add a short table of contents at the top.
  • Keep cross-references one level deep — link to a doc, not to a doc that only links onward.
  • Reference docs: never mirror upstream — record the delta. And never stamp a verification you did not run; an unverified claim is a documented one with a link, or it is left out.

Source: SKILL.md on GitHub

1 alert6mo3 checks · Risk SAFE
  • Gen Agent Trust Hub6mo

    The skill manages documentation review and generation by processing project files. While the core functionality is safe, the skill is susceptible to indirect prompt injection because it reads and analyzes untrusted documentation and project instructions (CLAUDE.md), which could contain malicious directions. No evidence of data exfiltration or remote code execution was found.

  • Socket6mo

    No alerts

  • Snyk6mo

    Risk: HIGH · 1 issue

Signed by skilld at 6e73347. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub yesterday.

Activeupdated 2 weeks ago
effort
high
Other metadata
argument-hint
[ (no args = context-aware assess) | --review | --update | --generate <target> | --session [--md <file>]] [--all | --staged | --unpushed]

README badge

README badge for nielsmadan/agentic-coding/doc