aindex waste --ingest agent-session.jsonl --source codex --json
Waste · Para: lead
Everyone has seen an agent read the same files again and again instead of moving forward.
Loop Killer watches session behavior and recommends a graph-aware next move when the agent drifts.
$ aindex waste --ingest agent-session.jsonl --source codex --json
loop_findings
wasted_reads
ignored_pack
ingested_events
recommended_next_call
contract_verifiedComando
Agent Loop Killer
Detect repeated reads, ignored packs and failing-test loops.
status detail=waste
loop_findingswasted_readsignored_packingested_eventsrecommended_next_callcontract_verified
Lo que el Cerebro hace automáticamente
El Cerebro dispara este workflow por ti.
The dashboard catches the exact behavior teams complain about: repeated reads and tests without progress.
end
When a session closes and outcomes are recorded.
—Tier requerido
Donde esto pasa a estar disponible comercialmente.
Ingests agent transcripts and flags repeated reads, ignored packs and failing-test loops with a next-call fix.
Foundation
Visible as a product workflow and pilot proof target; do not treat as fully shipped enterprise automation yet.
Foundation
Names the parity workflow and proof path without claiming full Sourcegraph-class depth.
- CLI waste
- CLI transcript ingest
- MCP status
- Dashboard waste table
- Pilot report
- Monitors
This workflow stays discoverable across setup, CLI, MCP status, docs, dashboard and pricing — never a hidden command.
Feature visibility checklist
Buyer-visible names covered by Agent Loop Killer.
This keeps grouped product names discoverable without claiming a separate maturity level from the underlying workflow.
Agent Loop Killer
Repeated reads, ignored packs and loop recovery actions are surfaced.
- Surface
- CLI waste, dashboard waste table, MCP status, monitors
- Buyer-safe status
- Foundation · Foundation
- Runtime
- preview only
Waste
Stop the agent when it starts paying the same token bill twice.
Repeated reads flagged