You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Independent AI agent security. For supported actions on configured routes, risky actions get a safe alternative or your approval before they run, so the task still gets done. Works with Cursor, Claude Code, Codex, Gemini CLI, Hermes CLI, and MCP tools.
Telegram approval inbox + risk-tiered auto-approval for your herdr AI agent fleet — blocked agents surface as Telegram cards with Approve/Deny buttons (zero-dependency Node.js)
Runnable, offline reproductions of one defect class in three agent frameworks: a value that is neither an explicit yes nor an explicit no reaches the execute path of a human-approval gate.
A purpose-built companion for awareness campaigns: registration, a before/after learning instrument, and answers drawn only from approved content, with a queue of the questions the programme has not yet answered. Arabic and English as equals.
A dashboard can flag the risk. The harder question is who’s allowed to act on it. BALLAST is a supply-chain control tower — energy and semiconductor scenarios — with approval-gated mitigation and replayable denied-vs-approved outcomes.
Real-time oversight of LLM agents — watch the agent execute, approve or edit every tool call, abort if it goes sideways. Third piece of the trilogy with Sentinel + Recourse.
Scheduled runs on a shared folder tree: per-project leases, a decision gate in front of irreversible actions, and a handover for whoever works next. Measured against the same runs without the protocol. Standard library only, 16 tests.
15 compétences Instagram fusionnées depuis deux paquets MIT, derrière un verrou de publication réellement appliqué : rien n'atteint un fournisseur sans approbation vivante liée au contenu exact. Complète mehdijabry/instagram-skills et mehdijabry/instagram-agent-skill.
A credit guard for paid generation APIs: runs start as requests with model and credit cost, a typed approval naming the count arms a single run, the guard relocks after, and a ledger records cost per run.