Borrowing it
Nothing to install: this file belongs to tatargabor/set-core. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/tatargabor/set-core/main/.claude/skills/cross-project-channel/SKILL.mdgit clone --depth 1 https://github.com/tatargabor/set-coreWrote 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/tatargabor/set-core/cross-project-channel)<a href="https://agentmods.dev/skills/tatargabor/set-core/cross-project-channel"><img src="https://agentmods.dev/badge/skills/tatargabor/set-core/cross-project-channel.svg" alt="Measured on agentmods" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 2 findings, 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 55 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.
- medium Excessive Agency · line 95 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
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 | $0.00082 | $0.03109 |
| Opus 5 | $0.00041 | $0.01554 |
| Sonnet 5 | $0.00016 | $0.00622 |
| Haiku 4.5 | $0.00008 | $0.00311 |
Grade A, and why
cross-project-channel 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.
How it starts
The opening of the file, as written. The whole thing — 180 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Cross-project agent channel
Moved out of
CLAUDE.mdon 2026-08-22 so it loads when it is needed rather than in every session. Nothing was cut — the text below is the section verbatim.
Cross-project agent channel — TEMPORARY (from 2026-07-24)
While set-core and a consumer project are being integrated, their two copilot sessions coordinate over a file channel, because no shared transport exists. Remove this section once the real transport ships.
Corrected 2026-07-29 — the stated reason was wrong, and wrong in the direction that closes
off a working option. This said .set-control is per-project, so send_message/get_inbox
"cannot cross a project boundary". Measured: the registry is global
(~/.config/set-core/projects.json, 38 entries), and _find_control_worktree()
(mcp-server/set_mcp_server.py:462) walks it and returns the first project that has a
.set-control directory — regardless of which project the server was started in. So both
sides would resolve to the same worktree; the project boundary was never the obstacle. The
real ones are that no .set-control exists anywhere (checked on both trees → both tools
return Error: No .set-control worktree found), and that the git-based ~15 s commit cycle is
rejected below. This is the proxy-instead-of-the-thing class applied to a code path: the
sentence described what the resolution ought to scope to, not what the function does.
DECISION 2026-07-24 — do not revive the git-based control sync for this. The existing
agent-messaging path (set-control-sync, .set-control worktree, ~15 s commit cycle) is
rejected as the cross-project channel: it caused problems in practice, and it carries
ephemeral messages where this coordination needs durable state. A live (non-git)
transport may be built later as its own piece of work; until then the file channel below is
the agreed mechanism.
Protocol — one file, one writer. Channel dir:
~/.local/share/set-core/channels/<consumer-slug>/ (the slug is runtime-derived; never
hard-code a consumer name here — see External Project Confidentiality below). Each side
appends only to its own file and reads the other's:
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 · 180 lines · 82 tokens per session scan A 46f4fb7119fc
cross-project-channel is a skill published in the GitHub repository tatargabor/set-core (35 stars, last pushed yesterday), licensed MIT. It adds 82 tokens to every session and 3,109 once invoked, about $0.0004 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 2026-09-06.
Other skills, from other repositories
atmos-custom-commands
Custom CLI commands: command definition in atmos.yaml, arguments, flags, shared native step types, when: conditions, output/UI steps, env vars, custom component types.
email-triage
Classify incoming emails into triage categories using content analysis, sender profile memory, and thread detection. Produces a category (URGENT, REPLYNEEDED, FYI, NEWSLETTER, NOTIFICATION, SPAM) with confidence and reasoning.
heartbeat
Periodic heartbeat check that reads HEARTBEAT.md from the workspace and executes any tasks listed there. Use for autonomous background monitoring, periodic maintenance, and proactive task execution. Triggered by a cron schedule.
hatch3r-handoff-prepare
Captures mid-work session state into a canonical handoff document at .hatch3r/handoffs/active/. Use when ending a session mid-work, switching tools, or after context-health Orange/Red.
hatch3r-cleanup
Repo-state cleanup pass — inventories accumulated hatch3r working state (plan files, workspace dirs, review telemetry, findings ledgers, learnings, handoffs, board staleness), asks once with per-category counts and estimated reclaim, dispatches each owned surface to its owning skill or CLI path, and prunes only the…
hatch3r-learn
Captures learnings from completed development sessions into reusable knowledge files for future consultation. Invoke manually, from board-pickup after PR merge, or with a specific issue number for targeted reflection.