agentleFS
iwesysiwesys/IWE/.claude/skills/iwe-update/SKILL.md

iwe-update

Skill0 starsChanged 25 days agoMITView on GitHub

Update IWE with change explanations. Agent calls update.sh, parses CHANGELOG,

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

11 rules. About 0 are common, 5 are distinctive, and 6 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 6

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Run update.sh in preview mode:

    line 31
  2. Distinctive:

    If "Everything is up to date" → notify the user and exit.

    line 43
  3. Distinctive:

    Before applying — check user customizations:

    line 77
  4. Distinctive:

    If extension files exist → check whether new extension points conflict with existing files.

    line 83
  5. Very specific:

    Read params.yaml. If the update adds new parameters → notify the user:

    line 86
  6. Very specific:

    Check whether .claude.md.base exists:

    line 90
  7. Distinctive:

    Does not exist → warn: "First update after migration.

    line 92
  8. Very specific:

    If the user approved (or --yes):

    line 96
  9. Very specific:

    If CLAUDE.md contains <<<<<<< markers → help resolve them:

    line 112
  10. Very specific:

    If the update introduced new parameters in the template's params.yaml → offer to add them to the user's params.yaml.

    line 118
  11. Distinctive:

    If the update introduced new extension points in protocols → notify:

    line 121

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

IWE Update (updater agent)

Role: R6 Coder. Single output: updated IWE + explanation of what changed. Trigger: "update IWE", "run update", /iwe-update, or the agent proposes after detecting a new version.

Algorithm

0. Extensions before

bash .claude/scripts/load-extensions.sh iwe-update before

Exit code 0 → read each path from the output in alphabetical order and execute instructions before the preview. Exit code 1 → no extensions, continue. Any other code or a required extension step failure → do not start the update; show the actual error cause.

1. Preview (--check)

Run update.sh in preview mode:

(cd "$IWE_TEMPLATE" && bash update.sh --check 2>&1)

Extract from the output:

  • Upstream version (current vs. new)
  • List of new files
  • List of updated files
  • Number of changed lines

If "Everything is up to date" → notify the user and exit.

2. CHANGELOG — what changed

CHANGELOG.md is no longer delivered locally (WP-401 Phase 6.1) — read via raw URL:

curl -fsSL "https://raw.githubusercontent.com/iwesys/IWE/main/CHANGELOG.md"

Find entries between the current and new version.

Explain to the user in plain language:

  • What was added and why (not a list of files, but the benefit)
  • What changed in behavior (could anything break)
  • What was fixed (bugs that may have affected the user)

Explanation format:

## IWE Update: v{old} → v{new}

**What's new:**
- [1–3 items in user language, not developer language]

**What changed:**
- [behavior the user may notice]

**Fixes:**
- [if any]

3. Compatibility analysis

Before applying — check user customizations:

3a. Extensions:

ls extensions/*.md 2>/dev/null

If extension files exist → check whether new extension points conflict with existing files.

3b. params.yaml: Read params.yaml. If the update adds new parameters → notify the user:

"A new parameter X appeared (default: Y). Do you want to change it?"

3c. CLAUDE.md: Check whether .claude.md.base exists:

  • Exists → 3-way merge will run automatically.
  • Does not exist → warn: "First update after migration. CLAUDE.md will be updated with the USER-SPACE section preserved. Review the result."

4. Applying the update

If the user approved (or --yes):

(cd "$IWE_TEMPLATE" && bash update.sh --yes 2>&1)

Extract from the output:

  • How many files were updated
  • Whether 3-way merge conflicts occurred
  • Whether variable substitution succeeded
  • Whether platform-space was updated
  • Memory files: lines NOT updated / NOT delivered and the summary Memory files not updated: N (…) — update.sh left these files as-is (they contain edits, their integrity cannot be verified, or the backup failed); line Memory files replaced: N (…) — replaced files, with prior versions in the named directory.

5. Post-update

5a. 3-way merge conflicts: If CLAUDE.md contains <<<<<<< markers → help resolve them:

  • Show each conflict.
  • Explain: "The platform changed X, and you changed Y. Which one do you want to keep?"
  • Apply the user's choice.

5b. New parameters: If the update introduced new parameters in the template's params.yaml → offer to add them to the user's params.yaml.

5c. New extension points: If the update introduced new extension points in protocols → notify:

"New extension points appeared: [list]. You can create extensions/X.md files to customize them."

5d. Recommendations: Based on the user's customizations (their extensions/, params.yaml, Pack) — suggest how to use the new capabilities.

5e. Extensions checks:

bash .claude/scripts/load-extensions.sh iwe-update checks

Exit code 0 → read and execute each file before the final report. Exit code 1 → no checks. Any other code or a check failure does not roll back already-applied files, but prevents declaring the update fully successful: show "update applied, user check failed", the exact cause, the path of the failed extension, and the list of already-executed checks files. Remediation: fix or temporarily disable that specific extension, reload iwe-update checks, and re-run all checks. Until they pass, the state is called applied_with_failed_extension_check, not a successful update.

6. Report

## Update complete

**Version:** v{old} → v{new}
**Files updated:** N
**Conflicts:** 0 / N (resolved)
**New capabilities:** [brief list]
**Memory files not updated:** N — [list from the "Memory files not updated" summary; for each — the reason from its "NOT updated …" line]
**Memory files replaced:** N (prior versions: [directory from summary])

Restart Claude Code to apply updates from memory/.

Memory files left as-is will not update themselves. For each one, show the user a diff command (diff …) and the "accept template version …" command from its line. Execute the accept command only after the user confirms there are no local edits in the file — it replaces the file with the template version, leaving the prior copy alongside it. If a line reports a backup or copy failure — resolve the cause (permissions, disk space) first and re-run update.sh.

7. Extensions after

bash .claude/scripts/load-extensions.sh iwe-update after

Exit code 0 → read and execute files after the report. Exit code 1 → no extensions. An after error does not revert the already-applied update, but must be shown as a separate warning with the extension name and cause.


Automatic update proposal

During Day Open (Step 5: overnight IWE check) — if update.sh --check shows available updates → add to "Requires attention":

"IWE update v{new} is available. Run /iwe-update to update."

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.