What does this skill do, and when should you use it?
wiki-fold rolls recent entries from wiki/log.md into a metadata page under wiki/folds/. Its summaries are extractive: outcomes and themes must be traceable to specific child entries, with no invented facts. Dry-run is the default and emits content to stdout without writing. Commit mode writes the fold and updates the index and log, while the skill explicitly excludes hierarchical fold-of-folds, automatic triggering, and semantic-tiling deduplication.
Reads the latest 2^k entries from wiki/log.md and records their dates, operations, titles, locations, and links; optionally reads up to 15 referenced pages; creates a deterministic fold ID and target path under wiki/folds/; checks for duplicate targets, insufficient batch size, child/table consistency, and verifiable numeric claims; emits dry-run content through Bash stdout; and in commit mode uses write/edit operations, wiki-lock.sh, and the repository's PostToolUse auto-commit workflow to update the fold page, wiki/index.md, and wiki/log.md.
- Obsidian users who want periodic rollups of a growing activity log into readable milestone pages.
- Researchers or engineers who need summaries with explicit source traceability and no unsupported synthesis.
- Claude Code users who need repeatable, structurally idempotent batches over a fixed number of log entries.
- Multi-writer vaults where concurrent fold-page writes require advisory locking.
How do you install this skill?
- Commit mode can trigger the PostToolUse auto-commit for wiki, .raw, or .vault-meta changes; confirm the staged scope and preserve a rollback point first.
- --force overwrites an existing page with the same fold ID; prefer duplicate detection and review extracted content and every numeric claim.
- The skill depends on Bash/GNU utilities, references/fold-template.md, wiki-lock.sh, and the expected vault structure; the supplied material does not prove all dependencies are available in the target environment.
- The publisher is not verified by the FollowSkills enterprise registry and should be treated as unknown.
- Shell / CLI
- Local filesystem
BashGit
There is no standalone installation procedure documented for this skill; it ships as one of 15 skills in claude-obsidian. Run: git clone https://github.com/AgriciDaniel/claude-obsidian && cd claude-obsidian && bash bin/setup-vault.sh. Then open the folder in Obsidian, open Claude Code in the same folder, and run /wiki for setup.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/AgriciDaniel/claude-obsidian.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/skills/wiki-fold" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- fold the log, dry-run k=3
- fold the log, commit k=3
Start with the default dry-run, for example: "fold the log, dry-run k=3". The skill reads the latest 2^3 entries and stops if the batch is short or the deterministic fold file already exists. After a clean dry-run, request: "fold the log, commit k=3". An explicit range can override k, and --force can overwrite an existing fold. Commit mode writes the fold page, adds its index link, and prepends a fold entry to wiki/log.md.
What are this skill's strengths and limitations?
- Dry-run produces no write-tool calls, making review possible before changes.
- Deterministic IDs provide structural idempotency for the same range and batch exponent.
- Extractive rules and count checks constrain unsupported summaries.
- Child entries remain untouched, and wiki-lock.sh addresses concurrent fold writes.
- It handles only flat folds over raw wiki/log.md entries; hierarchical fold-of-folds is out of scope.
- It refuses to silently process a partial batch when fewer than 2^k entries exist.
- It depends on Bash, filesystem access, repository scripts, and the repository's PostToolUse auto-commit behavior for commit mode.
- The supplied material references references/fold-template.md but does not include its contents, so the canonical final template cannot be independently checked here.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Wiki Log Folder this page | 55 · Use with care | ★ 16k | 1mo ago | MIT |
| Conversation Archivist | 49 · Use with care | ★ 16k | 1mo ago | MIT |
| Wiki Knowledge Companion | 49 · Use with care | ★ 16k | 1mo ago | MIT |
| Wiki Ingest for Obsidian | 49 · Use with care | ★ 16k | 1mo ago | MIT |
| Autonomous Research Filing | 52 · Use with care | ★ 16k | 1mo ago | MIT |
How did FollowSkills review this skill?
The skill explicitly defaults to dry-run with no writes, requires explicit user confirmation before commit, discloses the auto-commit hook, and specifies locking, duplicate detection, additive behavior, and reversal paths. Deductions apply because --force can overwrite an existing fold, commit mode creates external repository state changes, sensitive-data handling and dependency-integrity checks are not documented, and publisher identity is unverified.
The procedure, parameters, scope limits, self-checks, and failure messages cover short batches, duplicate files, missing pages, and lock conflicts. Deductions apply because no dedicated tests or reproducible evidence cover wiki-fold key paths, and execution depends on Bash, grep, head, the referenced template, wiki-lock semantics, and the surrounding vault environment; static calibration prevents a higher score.
Triggers, inputs, defaults, output modes, and explicit non-goals are well specified for human-invoked Obsidian log folding, and the core function does not depend on an overseas online service. Deductions apply because non-Unix portability is not addressed and there is no evidence of Chinese-language interaction or examples.
The skill has clear information architecture, progressive detail, parameters, examples, boundaries, invariants, concurrency guidance, a referenced canonical template, and repository MIT license context. Deductions apply because the skill itself lacks a clearly assigned maintainer, update path, independent version record, or changelog, and relies on other files for template and locking semantics.
The objective, deterministic naming, extractive constraints, count checks, and dry-run/commit outcomes are concrete and potentially directly usable. Deductions apply because no representative generated fold or key-path verification is provided, LLM extraction still requires review, and referenced dependencies may require additional setup; static calibration limits the score.
The supplied sources contain detailed operational rules, a canonical template, self-check requirements, and repository-level CI and test evidence. Deductions apply because the shown tests mainly cover other scripts and do not exercise wiki-fold parsing, extraction, locking, duplicate detection, or commit behavior; there is no third-party execution evidence or independent corroboration.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →