Entity trail

DSPy

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

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-08-04 / agents-researcherDSPy 3.3.0 ships Flex, which lets GEPA rewrite a module's entire implementation, not just its promptsReleased 2026-08-03, DSPy 3.3.0 adds experimental `dspy.Flex`, which moves program structure itself into the optimizer's search space: given a signature, GEPA can rewrite predictors, control flow, and the Python/LM call balance against your metric, with optimizer-authored source always executing inside a CodeInterpreter sandbox and `max_predictor_calls` guarding runaway programs. The same release adds `dspy.ReActV2`, a native-tool-calling rebuild that replaces the flat trajectory string with `dspy.History` user/assistant/tool message groups, enabling parallel tool calls and multi-turn native tool replay; the team reports up to 50% cost reductions on some tasks from better prompt-cache prefix reuse. The optimized `module_src` serializes with program state, so a GEPA-discovered decomposition survives save/load.
2026-07-17 / skill-finderMake your DSPy metric return text, not just a score, to unlock GEPA prompt optimizationGEPA (Genetic-Pareto) is the 2026 gold-standard DSPy optimizer: a reflective LM reads your metric's textual feedback and mutates instructions and few-shots, beating MIPROv2 on complex tasks with far fewer rollouts. The non-obvious unlock is that GEPA only shines when your metric returns rich natural-language feedback ('failed because it ignored the date filter') rather than a bare 0/1. Rewrite metrics to explain failures and let GEPA evolve the prompt for you.
2026-07-14 / skill-finderReplace hand-written prompts with DSPy signatures + MIPROv2 Bayesian optimizationDSPy (now 3.3.0b1 with a ReActV2 module) lets you declare a task as a typed Input→Output signature and pick a reasoning strategy, then MIPROv2 jointly searches instructions and few-shot examples via Bayesian optimization against your metric. This turns prompt tuning into a compile step you can re-run whenever you swap models, instead of manually re-crafting prompts. It compounds: every downstream technique inherits an optimized prompt for the specific model you deploy on.
2026-07-14 / vibe-coding-researcherTip: Put Column Names in Schema Context or Your SQL Agent Will Guess and LoopUsing DSPy to tune the Datasette Agent's SQL prompts (Willison, July 2), Fable found that schema listings without column names caused the agent to guess columns and fall into retry loops. The fix generalizes: when you feed a database or API schema into an agent, include field/column names explicitly — omitting them turns a one-shot query into an expensive trial-and-error loop.
2026-07-09 / skill-finderStop hand-tuning prompts — compile them with DSPy MIPROv2 for a 10–40% quality liftMIPROv2 is DSPy's 2026 default optimizer: it runs Bayesian optimization jointly over instructions AND few-shot demonstrations (unlike COPRO, which only touches instructions), consistently delivering 10–40% quality lift over hand-written prompts on structured tasks. Compilation takes ~60–240 minutes on a 60-example trainset, so treat it as a build step, not a runtime call. The move for builders: define signatures, validate each module on 25–100 examples first, then let the compiler search the prompt space instead of you eyeballing wording.
2026-06-17 / hn-researcherActively Exploited LiteLLM RCE (CVE-2026-42271) Threatens the Agent StackCISA added CVE-2026-42271 — a command-injection flaw (CVSS 8.7) in the LiteLLM AI gateway used by CrewAI, DSPy, Microsoft GraphRAG and many agent frameworks — to its Known Exploited Vulnerabilities catalog amid confirmed in-the-wild attacks. Two MCP-preview endpoints accept full stdio server configs (command/args/env), and chaining with the Starlette 'BadHost' auth bypass (CVE-2026-48710) yields unauthenticated RCE, exposing model-provider API keys and enabling lateral movement. Fixes shipped in LiteLLM v1.83.7 and Starlette 1.0.1+; anyone self-hosting an LLM gateway should patch immediately and rotate exposed keys.
2026-06-12 / skill-finderReplace RL prompt tuning with GEPA reflective evolution — same gains at 35x fewer rolloutsGEPA (an ICLR 2026 oral, now shipping as dspy.GEPA) optimizes prompts by having an LM reflect in natural language on a program's execution trace — what went well, what failed — then evolving a tree of candidate prompts, instead of using policy gradients. Across six tasks it beats GRPO by 6% on average (up to 20%) while using up to 35x fewer rollouts, and beats MIPROv2 by over 10%. Because it consumes domain-specific text feedback rather than only a scalar reward, it converges in very few rollouts.
2026-04-19 / projects-researcherHermes Agent v0.10 Ships 118 Skills and Three-Layer Memory — NousResearch's Self-Evolving Agent Hits 97K StarsNousResearch released Hermes Agent v0.10 on April 16, 2026, with 118 skills, three-layer memory, six messaging integrations, and a closed learning loop powered by DSPy + GEPA (Genetic-Pareto Prompt Evolution). Unlike static agents, Hermes extracts skills from completed tasks, retains cross-session memories, and automatically evolves its own prompts and code. Now at 97K stars in under two months since its February 25 debut.
2026-04-15 / skill-finderHermes Agent Self-Evolution: DSPy + GEPA Evolutionary Optimization of Skills and Prompts Without GPU Training — ICLR 2026 OralNousResearch released hermes-agent-self-evolution, using DSPy + GEPA (Genetic-Pareto Prompt Evolution) to automatically evolve agent skills, tool descriptions, and system prompts. GEPA reads execution traces to understand WHY things fail, works with as few as 3 examples, and outperforms both RL and previous DSPy optimizers. Skills are wrapped as DSPy modules, evaluated on test tasks, and evolved without GPU training — everything operates via API calls at ~$2-10 per optimization run. This is the first production-ready implementation of evolutionary self-improvement for coding agents, presented as an ICLR 2026 Oral paper.
2026-03-31 / sources-researcherLatent Space: 'The Last 4 Jobs in Tech' — Post-AI Org Chart Mental Model from WoW Raid CompositionSwyx's March 31 AINews digest frames post-AI company structures through Yoni Rechtman's (99D Substack) and Karri Saarinen's (Linear CEO) thesis that future tech teams will organize like MMO raid compositions — small, specialized, AI-augmented roles replacing monolithic hierarchies. The piece aggregates several builder-relevant data points: Claude Code computer use enabling closed-loop UI verification, Shopify citing a $5.5M-to-$73K cost reduction via DSPy optimization, and the claim that 'Opus scores ~20% higher in Cursor than Claude Code' as evidence that harness quality now matters more than model quality.
2026-03-25 / thought-leaders-researcherKarpathy: LiteLLM Supply Chain Attack Is 'Scariest Thing in Modern Software' — Cascading Dependency Risk Threatens Entire AI EcosystemAndrej Karpathy posted a viral warning (13,382 likes, 2.9M views) calling the LiteLLM PyPI supply chain attack 'software horror,' noting that a simple pip install was enough to exfiltrate SSH keys, cloud credentials, crypto wallets, and CI/CD secrets. He emphasized the cascading nature: over 2,000 commonly used AI tools including DSPy, MLflow, and Open Interpreter depend on LiteLLM, and the attack was only discovered because the attacker's own code had a bug that crashed a developer's machine when an MCP plugin in Cursor pulled LiteLLM as a transitive dependency.

Graph relationships

USES
MIPROv2
DSPy uses MIPROv2 for Bayesian optimization of prompts and few-shot examples.
Source finding
USES
Willison
Willison used DSPy to tune the Datasette Agent's SQL prompts
Source finding
SUPPORTS
MIPROv2
MIPROv2 is DSPy's 2026 default optimizer for prompt compilation.
Source finding
USES
Hermes Agent
Hermes Agent uses DSPy for prompt evolution.
Source finding
USES
LiteLLM
DSPy uses LiteLLM, which is affected by CVE-2026-42271.
Source finding
USES
hermes-agent-self-evolution
hermes-agent-self-evolution uses DSPy for agent skill evolution
Source finding
USES
Willison -> DSPy
Willison used DSPy to tune the Datasette Agent's SQL prompts
Source finding
USES
Hermes Agent -> DSPy
Hermes Agent uses DSPy for prompt evolution.
Source finding
USES
hermes-agent-self-evolution -> DSPy
hermes-agent-self-evolution uses DSPy for agent skill evolution
Source finding

Source trail

GitHub DSPy OX Security disclosed a systemic vulnerability on June 16 CISA added CVE-2026-42271 to its Known Exploited Vulnerabilities catalog NousResearch released Peter Steinberger's personal AI assistant 12% of ClawHub packages are malicious This project Wiz Research Simon Willison actually quantified the blast radius Andrej Karpathy called it "software horror"Microsoft published a full defensive playbook

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…