oxygen-unibox
Skill2 starsChanged 2 months agoMITView on GitHub
Triage OXYGEN Messages and Unibox conversations, prepare grounded replies and send only within the user's specified account and thread scope.
In this repository54 other files it gives its agents.
- AGENTS.mdAGENTS.md
- context/AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- context/CLAUDE.mdCLAUDE.md
- brand-reviewSkill
- brand-systemSkill
- capture-contextSkill
- content-strategySkill
- crm-presync-checkSkill
- flowchartSkill
- graphics-designerSkill
- gtm-startSkill
- launch-videoSkill
- lead-magnet-creatorSkill
- lead-sourcingSkill
- linkedin-copywriterSkill
- long-formSkill
- manim-videoSkill
- newsletter-writerSkill
- outbound-copywriterSkill
- oxygen-dedupeSkill
- oxygen-deliverabilitySkill
- oxygen-diagnosticsSkill
- oxygen-email-infraSkill
- oxygen-gtmSkill
- oxygen-knowledgeSkill
- oxygen-linkedin-marketingSkill
- oxygen-quickstartSkill
- oxygen-recipesSkill
- oxygen-sequencer-enrollSkill
- oxygen-sequencerSkill
- oxygen-table-tidySkill
- oxygen-workflow-authoringSkill
- personalize-ai-columnSkill
- qmdSkill
- repurpose-contentSkill
- researcherSkill
- setup-workspaceSkill
- shorts-audioSkill
- shorts-cutSkill
- shorts-editSkill
- shorts-motionSkill
- shorts-qaSkill
- signals-sweepSkill
- spin-tagsSkill
- split-test-designerSkill
- video-useSkill
- voice-calibrationSkill
- week-postsSkill
- x-copywriterSkill
- youtube-descriptionSkill
- youtube-publisherSkill
- youtube-scriptSkill
- youtube-thumbnailSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 8
Read the runtime contract.
- Distinctive: line 12
Read the full relevant conversation and current Knowledge/CRM context before classifying or drafting.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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.
Files
Work the existing conversation
Read the runtime contract. Messages owns existing email, LinkedIn and WhatsApp threads, plus direct email without cadence. New LinkedIn conversations and outreach programs use Sequences.
Discover with oxygen capabilities search "triage unanswered inbox threads" --json, then hydrate the selected inbox/messages/reviews command. Query a bounded slice by channel, time, status, participant or shared Tag. An unfiltered inbox includes answered threads; use the documented unanswered filter for a reply queue.
Read the full relevant conversation and current Knowledge/CRM context before classifying or drafting. Preserve provider thread IDs, participants, sending account and delivery state. Use the actual message-review surface when appropriate.
LinkedIn visibility is sender-scoped. Inspect the account's inbox scope if expected threads are missing. Do not interpret an empty restricted view as the entire inbox or opt an account into full ingestion merely to find more messages.
For a reply, prepare the exact content, thread, account and recipients. Send only within the user's explicit authorization. For tags/archive/bulk changes, expose the actual count and scope and preserve unrelated conversations.
A message timeout is not proof of non-delivery. Inspect returned message/run provenance before any retry. A public comment on a post belongs to Posts/Publishing, not Unibox.
Return a bounded triage summary, draft or verified message links, classifications and unresolved threads. Do not file feedback or forward conversation/transcript content elsewhere merely because an error occurred.
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
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.

