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
{{ message }}
Repository navigation
insights: the self-removed-credit finding still shows a saving context-guru did not make #584
/context-guru:insights still shows a dollar figure for what the account removed by itself: the self-removed-credit finding. The repo owner decided in #583 that a saving context-guru did not make is not shown as ours. #583 removed the self_removed row from the levers table. This finding is the one place left.
What we found
context-guru-plugin/scripts/insights.py:856-880 reads tools.self_removed and adds the finding self-removed-credit (severity ok): "You already removed N capability(ies), and it worked ... $X avoided over D days".
It never goes into a saved total. The findings list ranks it by its own dollar figure (usd_window, usd_month), so it can appear high in the list.
It sets the facts self_removed_items and self_removed_usd.
The repo owner, on #583: "what Claude removes is not our concern, remove the row." In the levers table, self_removed is gone.
Open question for the assignee
The code comment calls this "a saving the product caused": a server stops being declared, for example after /context-guru:remove-mcp-and-skills recommended removing it. If the removal came from our own recommendation, it is arguably our saving, and #537 already measures "savings since removal" for that command. Please choose one, and confirm it with the repo owner:
Activity