Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add Ancienttwo/repo-harness --skill obsidian-memorygit clone --depth 1 https://github.com/Ancienttwo/repo-harnessWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/ancienttwo/repo-harness/obsidian-memory)<a href="https://agentmods.dev/skills/ancienttwo/repo-harness/obsidian-memory"><img src="https://agentmods.dev/badge/skills/ancienttwo/repo-harness/obsidian-memory/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/ancienttwo/repo-harness/obsidian-memory"><img src="https://agentmods.dev/badge/skills/ancienttwo/repo-harness/obsidian-memory.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Rogue Agent · line 88 Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.Fix: Remove any persistence mechanisms (cron jobs, startup scripts, state files). Skills should not maintain state across sessions without explicit user consent.
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00212 | $0.02280 |
| Opus 5 | $0.00106 | $0.01140 |
| Sonnet 5 | $0.00042 | $0.00456 |
| Haiku 4.5 | $0.00021 | $0.00228 |
Grade A, and why
obsidian-memory scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 91 lines — stays where its author put it; the contents beside it link to each section on GitHub.
obsidian-memory — cross-project long-term memory (Obsidian brain vault)
Outcome Contract
- Outcome: recall relevant background from the vault before a task; after the task, persist the conclusions that will still be useful later into the matching project sub-vault, and keep the index current.
- Done when: the recalled notes have been folded into the current context as leads (after re-verification against current state), or the added/updated notes are on disk and the sub-vault
index.mdis in sync. - Authority model (inviolable): the in-repo artifacts (
tasks/,docs/,MEMORY.md, the code) are each project's source of truth; the vault is only a cross-project aggregation/projection layer, and the direction is always repo → brain. When the vault conflicts with current state, current state wins, and the vault gets corrected on the spot. - Two readers: the vault is read by agents and by the user. A note must be prose a person can read and learn from directly — complete sentences that state the why and the tradeoff, not agent shorthand and not piles of raw logs;
index.mdis the human reading entrypoint. - Boundary: this skill may only be invoked explicitly by the model or the user. Hooks never execute it — the hook layer may at most emit a
[BrainPromote]-style advisory string, and never reads or writes vault state.
Vault resolution (fail-closed, and the vault itself is optional)
- Read
brainRootfrom~/.repo-harness/config.json. - Not configured, or the path does not exist → stop and say so. Do not scan the disk to guess a vault, and do not create a vault root on the fly.
- The project sub-vault is
<brainRoot>/<project-slug>/;<project-slug>is the repo directory name or a name the user gives.
Having no brainRoot configured is a legitimate steady state, not a defect awaiting repair. Unconfigured simply means this machine does not use the vault layer: the in-repo artifacts remain a complete, authoritative memory surface, and at closeout the conclusions go into the existing slots such as tasks/lessons.md and docs/researches/. Do not create a vault just so this skill can run. Only when the user explicitly wants the vault layer enabled, point them at repo-harness install --brain-root <path> or repo-harness update --brain-root <path>.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago Changed 7f6adb89a7e7
- 9d ago First seen · 91 lines · 212 tokens per session scan A 434ceb8d391f
obsidian-memory is a skill published in the GitHub repository Ancienttwo/repo-harness (431 stars, last pushed today), licensed MIT. It adds 212 tokens to every session and 2,280 once invoked, about $0.0011 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
codebase-memory-mcp
Resolve canonical Git checkouts, index and verify codebase-memory-mcp graphs through the guarded CLI, and safely audit duplicate worktree caches. Use when a user mentions codebase memory MCP, searchgraph, tracepath, indexrepository, worktree indexes, or oversized graph caches.
mastermind-project-history
Retrieve and reason from durable project decisions, failed approaches, audits, reports, and lessons without treating provenance, search rank, or user approval as technical proof. Use when asking why a design exists, whether an approach was tried, what supersedes an older decision, what prior evidence should constrain…
maintain-codex-wiki
Maintain a review-first Markdown knowledge base for Codex practices with source provenance, engineering capture, citation-aware queries, explicit archive and promotion, and deterministic linting. Use when asked to capture a durable engineering lesson, ingest Codex research, query or archive what the repository knows…
context-compactor
Scans codebase directories and generates a high-level, token-efficient structural overview of all classes, functions, and docstrings, omitting implementation details.
workspace-mirror
Sync and backup Antigravity IDE sessions, transcripts, SQLite database metadata, and brain artifacts locally into a structured workspace directory (.gemini-local/) with a rich interactive viewer.html page. Use this skill whenever the user mentions backups, workspace syncing, fswatch real-time mirroring, viewing…
capture-to-vault
Use during /research, /scope (or any recipe) to write or update a vault note via dossier-mcp. Agent drafts slug + frontmatter + body + citations, writes the note, and surfaces what was captured so the user can redirect afterward. Supports create (new note) and update (refine existing note) modes.