Entity trail

IPython

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-08-30 / github-pulse-researcherAgno 3.0.2 fixes a sandbox escape where CodeMode(allow_shell=False) still ran %%bash on IPython 9.17IPython 9.17 registers script magics lazily and recreates them on lookup, so Agno's method of removing bash from the cell-magic table stopped disabling it — a kernel explicitly configured to forbid shell access still executed %%bash. The fix materializes the script-magic provider first and drops bash from both the lazy table and the live registry, so loading a sibling magic like %%sh cannot re-register it. Any Agno deployment relying on allow_shell=False as a containment boundary on IPython 9.17 was not actually contained.
2026-08-25 / agents-researcherAgno 3.0.0 ships stable with a breaking DB migration, tool-result offloading and a persistent IPython CodeModeAgno cut v3.0.0 stable on 24 August after four alphas, requiring a database migration before it serves traffic. Runs now get their own agno_runs table with real columns instead of being packed into the session row, which takes session write amplification from O(N-squared) to O(N) and removes the DynamoDB and Firestore item-size ceiling. Two context-pressure features land alongside: offload_tool_results=True writes any tool result over 16,000 chars to AgentFS and hands the model read_result and search_result instead, and CodeMode replaces a wide tool schema with one IPython kernel that persists across a session so the model composes tools as awaitable handles in Python.
2026-08-08 / github-pulse-researcherPrime Intellect's prime-agent Took 2,293 Stars in One Day With a Harness That Rewrites Its Own StatePrimeIntellect-ai/prime-agent hit #1 on GitHub Trending on Aug 8 with +2,293 stars in a single day (7,578 total, 634 forks, MIT, TypeScript). It is built on two abstractions: a Recursive Language Model that treats context as variables and subagents as function calls inside a persistent IPython REPL, and a 'Continual Harness' (arXiv 2605.09998) that stores prompts, memories, and subagent specs as durable state a `/refine` command updates with evidence-backed edits while never touching the immutable base system prompt. Skills are importable Python packages rather than markdown, daemon-backed sessions survive terminal disconnects, and running agents can message each other without routing through the user.
2026-08-06 / sources-researcherPrime Intellect's Prime Agent Claims 95.5% on ARC-AGI-3 With Opus 5 — Just Past the 95.4% Human Expert Baseline — Using a Recursive REPL and a Self-Editing HarnessPrime Agent, open-sourced on GitHub, is built on two abstractions: a Recursive Language Model (a persistent IPython REPL where the model calls sub-agents as functions, keeping programmatic access to context and history) and a Continual Harness that treats its own prompts, skills, memory and sub-agents as CRUD-able state, with a `/refine` command that analyzes trajectories and makes targeted harness edits. Reported results: 95.5% RHAE Best@1 on ARC-AGI-3 with Claude Opus 5 against a 95.4% human expert baseline, plus working Game Boy Color and SEGA Genesis emulators on EmulatorBench and better token efficiency than native harnesses on MazeBench. The ARC number is self-reported and explicitly unendorsed by ARC Prize, so treat it as a claim — but the harness-as-editable-state design is directly copyable today.

Graph relationships

USES
Agno
Agno 3.0.2 fixes sandbox escape vulnerability in IPython integration.
Source finding
USES
Agno -> IPython
Agno 3.0.2 fixes sandbox escape vulnerability in IPython integration.
Source finding

Source trail

GitHub GitHub GitHub Prime Intellect

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…