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 Abilityai/cornelius --skill ref-reconcilegit clone --depth 1 https://github.com/Abilityai/corneliusWrote 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/abilityai/cornelius/ref-reconcile)<a href="https://agentmods.dev/skills/abilityai/cornelius/ref-reconcile"><img src="https://agentmods.dev/badge/skills/abilityai/cornelius/ref-reconcile.svg" alt="Measured on agentmods" height="20"></a>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 | $0.00094 | $0.02035 |
| Opus 5 | $0.00047 | $0.01018 |
| Sonnet 5 | $0.00019 | $0.00407 |
| Haiku 4.5 | $0.00009 | $0.00203 |
Grade A, and why
ref-reconcile 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 3d 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 — 93 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Ref Reconcile
ℹ️ First, set expectations: before anything else, print one short line with this skill's version and its most recent change — the top entry of
metadata.changelogabove — e.g.ref-reconcile v1.0 — recent: initial version. Then proceed.
The act on top of ref-audit's report. ref-audit senses (read-only, autonomous); ref-reconcile fixes (human-gated). It protects the property that makes a reference scope a source of truth: one canonical note per real-world entity, with consistent bidirectional links and no dangling references.
Read first: resources/layered-brains/COMPANY-BRAIN-SCHEMA.md → "Canonical source of truth" + "The update model"; resources/layered-brains/REFERENCE-SCOPE-SCHEMA.md.
Composes
/ref-audit— the sense/detection pass (do not re-implement the scan here).
State Dependencies
| Source | Location | Read | Write | Description |
|---|---|---|---|---|
| Company family | Brain/Company/*/*.md |
✓ | ✓ (gated) | Entity notes reconciled |
| Audit report | resources/company-brain/reports/ |
✓ | Consumed for detection | |
| LBS index | resources/local-brain-search/data/ |
✓ | ✓ (reindex) | Rebuilt after any write |
Process
Step 1 — SENSE (auto)
Detection has two sources — don't attribute to ref-audit what it doesn't compute:
From /ref-audit (structural, single-authored there — do not re-implement): dangling links, orphan entities, duplicate candidates, plus the frontmatter/provenance/enum/partition/as_of violations. Invoke it (or read its latest report).
In this skill (relational, needs the references-vs-relations judgment ref-audit deliberately avoids):
- Conflicts — two notes asserting contradictory facts about one entity. Read the duplicate/near-name clusters
ref-auditflags and compare their bodies. - Bidirectional-link asymmetry — but only for relation links, never reference links. A
## Relationsentry that names a mutual tie (client: [[B]],Partner of: [[B]],Engagement: [[B]]) should reciprocate; a directional reference (built on [[Trinity]],competitor of [[X]],runs on [[Y]]) must not be forced to back-link.ref-auditreports the family CLEAN despite dozens of such directional links precisely because a blanket A→B⇒B→A check would be almost all false positives — so scope the reciprocity check to relation verbs only:
(The old space-only, capitalized set —cd Brain/Company # Mutual-tie relation links whose reciprocal is worth checking. # Case-insensitive (-i) and connector-agnostic: matches BOTH "Client of [[..]]" and the colon # form "client: [[..]]" the migrated notes actually use. The verb must LEAD the label (~12 chars) # so prose bullets that merely contain a verb word (e.g. "…problem owners…") aren't flagged. # No \b — BSD/macOS grep does not support it. TUNE THE VERB SET to the scope's real vocabulary. grep -rniE '^- [^[]{0,12}(client|partner|vendor|counsel|owner|owns|engagement|collaborat|advis)[^[]*\[\[' */*.mdClient of|Vendor to|…— matched only 4 of the family's ~28 mutual relations; the family is predominantly lowercase colon-form.) For each match, confirm the named target carries the reciprocal relation; if not, it's a repair candidate (Step 2).
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.
- 3d ago First seen · 93 lines · 94 tokens per session scan A cb41a604228f
ref-reconcile is a skill published in the GitHub repository Abilityai/cornelius (106 stars, last pushed 14d ago), licensed MIT. It adds 94 tokens to every session and 2,035 once invoked, about $0.0005 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 2026-09-03.
Other skills, from other repositories
media-ingest
Ingest video, audio, PDF, book, screenshot, and GitHub repo content into the brain. Multi-format handling with entity extraction and backlink propagation. Covers video-ingest, youtube-ingest, and book-ingest subtypes.
mem0-oss-to-platform
Plan and then execute a migration of a project from the mem0 open-source / self-hosted SDK (the local Memory class) to the mem0 Platform / hosted / managed SDK (the MemoryClient class). Use this whenever a developer wants to move, switch, or migrate their mem0 usage off OSS/self-hosted to the hosted API — e.g.…
Cortex
Operate Cortex, the LifeOS memory system — the typed Knowledge Archive (People, Companies, Ideas, Research with typed related: links) plus recall of prior work sessions, ISAs, and conversations. Search, add, harvest, develop, ingest, distill, graph-navigate, recall. USE WHEN cortex, knowledge, knowledge base, search…
memory
Use when the user asks to remember, recall, forget, update, search, or inspect durable OpenSquilla memory, including profile facts in USER.md and long-term notes in MEMORY.md or memory//.md.
ha-data-stores
Map of Hope Agent's local data stores and safe read-only query workflow. Use when the user asks where Hope Agent stores data, wants to inspect sessions/messages/memory/logs/background jobs/knowledge indexes/settings, asks the model to query local app data, or debugging requires checking persisted state. Trigger…
establishing-project-context
Use when the user asks to establish shared project language, or project work exposes a conflicting, renamed, or deprecated domain term that needs active semantic modeling. Routine small tasks stay on the fast path.