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 agentmods add skills/roynertr/cobie-mcp/applying-updatesnpx skills add roynertr/cobie-mcp --skill applying-updatesgit clone --depth 1 https://github.com/roynertr/cobie-mcpWhat 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 | $0.00016 | $0.00333 |
| Opus 5 | $0.00008 | $0.00167 |
| Sonnet 5 | $0.00003 | $0.00067 |
| Haiku 4.5 | $0.00002 | $0.00033 |
Grade A, and why
applying-updates 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 yesterday.
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.
What it actually says
Applying Updates
Inputs assumed
- pending_preview_id (required; apply blocked if none)
- active_project (excel_path resolved from project)
- pending_operation.draft (the draft to commit)
Tool sequence
- Verify pending_preview_id exists
- Verify confirm_token === "CONFIRM_APPLY" or user typed "CONFIRM APPLY"
- commit_changes (update_cobie or capture_installation with dry_run=false)
Output template
- Summary: "Applied N changes."
- Bullets: summary of changes (max 5)
- CTA: "Validation complete. Type 'validate' to re-check."
- Optional: "Type 'show details' to see full diffs"
Forbidden actions
- Never call update_cobie without confirm_token CONFIRM_APPLY
- Never apply when pending_preview_id is missing
- Never skip the confirmation gate
Tool usage rules
- Require pending_preview_id; reject apply if missing
- Require confirm_token "CONFIRM_APPLY" from UI or "CONFIRM APPLY" from user message
- Call commit_changes with draft from pending_operation
Confirmation gates
- Require explicit confirm token before any write
- UI button sends confirm_token: "CONFIRM_APPLY"
- Or user must type "CONFIRM APPLY" (case-insensitive)
Output formatting
Return human summary + bullet list (max 5) + CTA. Do not output raw JSON unless asked.
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.
- yesterday First seen · 42 lines · 16 tokens per session scan A 0f5cd1832504
applying-updates is a skill published in the GitHub repository roynertr/cobie-mcp (1 stars, last pushed 29d ago), licensed MIT. It adds 16 tokens to every session and 333 once invoked, about $0.0001 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 2026-08-31.
Other skills, from other repositories
scrape-batch
Extract many known URLs in one polite, rate-limited pass. Use when the user hands over a list of links, a set of search hits to read in full, or asks to "scrape these pages" / "pull the content from all of them". Drives extract(action="batch"), which fans out with per-domain rate limiting and returns partial results…
memory-commit
Use when the user explicitly says "remember this", "save this", "ghi nho", "luu lai", "save for next time", or otherwise asks to persist the immediately preceding context. Captures with the appropriate contexttype (decision, preference, fact, skill, task, conversation) so future sessions can retrieve it accurately.
recall-context
Use at session start, before significant decisions, or when a new task references a known project to recall mnemo memories matching the current working directory, recently edited files, or topic keywords. Helps maintain continuity across sessions and avoid redoing past research.
session-handoff
End-of-session knowledge capture — decisions, preferences, corrections, conventions, open questions.
octave-compression
Workflow for transforming prose into semantic OCTAVE structures. Covers tier selection, transformation phases, loss accounting, and decision rules. REQUIRES octave-literacy.
query-superset
Run bounded SELECT queries through an Apache Superset SQL Lab connection on Windows. Use when a user wants an Agent to configure a supported Superset instance, test connectivity, inspect tables or columns, execute SELECT queries, export small results, preserve query evidence, or diagnose authentication, permission…