djalmajr/skills

aim-write

Manually write a durable page to an ai-memory instance via the MCP (memory_write_page) — for decisions, rules, gotchas, schemas/contracts, or operational constraints the user explicitly wants remembered.

First seen May 28, 2026

Installation

$ npx skills add djalmajr/skills --skill aim-write

Summary

  • Manually write a durable page to an ai-memory instance via the MCP (memory_write_page) — for decisions, rules, gotchas, schemas/contracts, or operational constraints the user explicitly wants remembered.
  • Use when the user says 'remember this', 'save this decision', or 'annotate X'.
  • Not for routine notes (auto-capture handles those).

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from djalmajr/skills · top by installs.

npx skills add djalmajr/skills

Browse all from djalmajr/skills

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Repository health

Stars 4
Default branch main
Open issues 0
Status Active

Skill metadata

Parsed from SKILL.md frontmatter.

More metadata
short-description
Manual durable write to ai-memory

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 2,763 B
  • docs SUMMARY.md 353 B

History

  1. First seen on skills.sh
  2. First recorded snapshot · 61 installs

SKILL.md

aim-write

Persist a durable wiki page. Use only for canonical knowledge the user explicitly wants remembered — routine session notes are captured automatically by the hooks; do not hand-write those, and do not use a handoff for a permanent annotation.

Steps

  1. Pick the instance. From the repo's MCP config. If multiple ai-memory servers are

configured, ask which endpoint to write to.

  1. Pick workspace/project. Default to the repo's .ai-memory.toml; confirm if writing

elsewhere.

  1. Choose a path + kind. Use folder conventions so kind is derived and recall stays clean:

- decisions/ADR-NNNN-<slug>.md → decision - rules/<slug>.md → rule - gotchas/<slug>.md → gotcha - otherwise a topical path (e.g. runbooks/<slug>.md, <area>/<slug>.md). You can also set kind explicitly via frontmatter nodetype / the tool arg.

The rules/ prefix matters — use the underscore. Pages under rules/ (and pinned slots/) are ai-memory's highest-signal tier: they're surfaced verbatim in memorybriefing / memoryexplore / the dashboard, keyed by path GLOB 'rules/*'. A plain rules/ folder (no underscore) is just a normal page — it gets none of that and is only found by explicit query/read. So write durable rules to _rules/, never rules/.

Shared / cross-project rules (e.g. global agent conventions reused across every repo) belong in a dedicated scope — the reserved default/global scope — under rules/ there (write with scope: "global"). Recall is per-(workspace, project), so other projects reach them cross-scope via memoryquery scopes:[…] / memoryread_page (see aim-query), not automatically.

  1. Write the body as markdown with frontmatter (title, node_type, tags, optional

source:). Link related pages with [[path.md]] (paths carry .md; links resolve by exact path). Never put live secrets/credentials in a page — redact (<see secret manager>).

  1. Call memorywritepage (or the equivalent admin write) with workspace/project/path/body.
  2. Confirm the page is recallable (a quick aim-query for its topic).

Keep pages focused and distilled — one fact/decision/gotcha per page reads and recalls best.