Skip to content
Back to skills

Aid Update Document

ASecurity

Revise or extend a document that already exists, in one pass. Use this skill when a document has drifted from the code, is missing a section, or needs a correction -- and you want the change drafted against the current source rather than written from memory. It reads the existing document first, then edits it, grounded in and accuracy-checked against the Knowledge Base (`.aid/knowledge/`) and the project source. It resolves nothing: it drafts the change and shows you a diff, and only after yo...

  • 6 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added November 7, 2026
toolsshellbashgit

Works with

  • terminal

Security analysis

A100/100

Scanned November 7, 2026

npx -y skills add AndreVianna/aid-methodology --skill aid-update-document --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Aid Update Document?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Aid Update Document
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/andrevianna-aid-update-document/badge)](https://www.skillsdirectory.com/skills/andrevianna-aid-update-document)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: aid-update-document
description: >
  Revise or extend a document that already exists, in one pass. Use this skill when a
  document has drifted from the code, is missing a section, or needs a correction -- and you
  want the change drafted against the current source rather than written from memory. It
  reads the existing document first, then edits it, grounded in and accuracy-checked against
  the Knowledge Base (`.aid/knowledge/`) and the project source. It resolves nothing: it
  drafts the change and shows you a diff, and only after you approve is it written back. It
  never writes into `.aid/knowledge/` -- that is `/aid-update-kb`'s territory.
allowed-tools: Read Glob Grep Bash Shell Terminal Write Edit Agent
argument-hint: "<document + change> -- which existing document to update, and how"
---

# Update Document (update it now, resolve nothing)

`/aid-update-document` updates an **existing** document -- the sibling of
`/aid-create-document` for edits rather than new authoring.
Same collapse shape, same producer/verifier, same KB boundary; the difference
is it **locates and reads the existing document first**, edits it, and shows a **diff** at
the human gate (overwrite care is central).

- **Boundary vs the KB:** never writes `.aid/knowledge/`; KB edits route to `/aid-update-kb`.
- **Not a numbered pipeline phase**; does not route to `/aid-execute`.
- **Genre structures:** `.aid/runtime/templates/shortcut-scaffolding/document.md`.

State machine: **INTAKE -> AUTHOR -> VERIFY (loop) -> PRESENT [human gate, diff] -> WRITE
(on approval) -> DONE**. Print the `[State: NAME] -- {purpose}` entry line on each state.

---

## State: INTAKE

1. **Require a target document + a change.** Empty argument -> ask one bootstrapping
   question ("Which document should I update, and what change?") and wait.
2. **Locate + read the existing document.** If it cannot be resolved, ask (or, if the user
   meant a new doc, suggest `/aid-create-document`).
3. **Pick the path** (fast if the document + change are clear; guided otherwise) and
   **classify complexity** -> `aid-tech-writer` model/effort (sonnet/medium default; opus/high
   for a large rewrite). Verifier tier >= producer.
4. **Consult the Work Initiation Gate, then allocate the work folder + STATE.** First run
   the gate (`.aid/runtime/templates/work-initiation-gate.md`):
   `bash .aid/runtime/scripts/works/enumerate-works.sh` (main tree + every git worktree).
   Empty -> allocate, no prompt. Works exist -> ask new-vs-continuation; on **continuation**
   route to the chosen work's resume door and STOP (allocate nothing); on **new work**:
   create and enter the worktree per the gate's `§ 3a` step 2
   (`worktree-lifecycle.sh create <work-id> <name>`, STOP on a non-zero exit or empty path,
   else enter the resolved path), **then** allocate (`initiator: aid-update-document`;
   `phase` not driven).
5. **Read the design seed, if present.** If `.aid/design/document.md` exists, read it as
   prior context before drafting; it is an input, never a substitute, and is not modified
   by this run.

**Advance:** AUTHOR.

---

## State: AUTHOR

Dispatch **`aid-tech-writer`** (clean context, tiered) to produce the **revised** document
(a draft in the work folder, not yet written back), grounded in and accurate to the KB +
project source, preserving the document's existing genre structure.

**Advance:** VERIFY.

---

## State: VERIFY

Same as `/aid-create-document`: mechanical grounding check + a clean-context **`aid-reviewer`**
adversarial check (accurate, complete, no fabrication, structure preserved) -> `grade.sh`
-> loop on failure (3-cycle circuit-breaker -> IMPEDIMENT + `lifecycle: Blocked`).

**Advance:** PRESENT.

---

## State: PRESENT  (hard stop -- human final say)

Set `lifecycle: Paused-Awaiting-Input`. Present the revised document **as a diff against
the current file** + the target path. Await approval. Never writes `.aid/knowledge/`.

**Advance:** WRITE on approval; else DONE (draft kept in the work folder).

---

## State: WRITE  (only on approval)

Write the revision back to the existing document (the diff was already reviewed at PRESENT).
Then optionally print handoffs (`/aid-update-kb`, `/aid-create*`, ...).

**Advance:** DONE.

---

## State: DONE

Set `lifecycle: Completed`, `updated` now, append a `## Lifecycle History` row.

---

## Constraints

- **Resolves nothing**; edits + writes back on approval.
- **Reads the existing document first**; shows a **diff**; never silently overwrites.
- **Grounded + accurate** to KB + source; **NEVER writes `.aid/knowledge/`**.
- **Clean context**; **verification always a sub-agent dispatch** (`aid-reviewer`).
- **Tracking:** write STATE `lifecycle` at every transition.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.