EdTech Tracker

← Coding lens

system prompt

technique judged 98% likely a named technique or architecture

First seen
2024-06-26
Independent sources
18
Articles
95
Last 4 weeks vs prior
0 vs 41

See how it has spread across media tiers

Mentions over time

Articles per week (by publish date) using the term, in any spelling.
014.5292024-W262026-W112026-W41
Table view
WeekArticlesSources
2024-W2611
2024-W3311
2024-W4911
2025-W0311
2025-W2111
2025-W2311
2025-W3111
2026-W0122
2026-W0211
2026-W0322
2026-W0421
2026-W0811
2026-W0961
2026-W11191
2026-W1911
2026-W2111
2026-W2422
2026-W2522
2026-W2711
2026-W3111
2026-W3211
2026-W3611
2026-W3722
2026-W3822
2026-W3921
2026-W40295
2026-W41102

In their words

A single call to the model can carry: the user's message tool results (files, web pages, search results, API responses) the arguments the model gave a tool on the previous turn tool descriptions, which the agent framework sends on every call the system prompt Any of these can carry an instruction.

every 3 hours) — not * * * * * Nodes: 1) Fetch sources (RSS/HTTP/search) → normalize to [{title,url,snippet}] 2) Filter: drop items seen in last N days (static data / datastore) 3) LLM summarize with fixed system prompt + max tokens 4) IF empty → stop (no notify) 5) Write artifact (git path or object key with date) 6) Notify summary + link to n8n execution On error: retry 2x with backoff → alert w

join ( parts ) Part 3: Agent System Prompts Supervisor Agent # prompting/agent_prompts.py SUPERVISOR_PROMPT = """ You are a Supervisor Agent managing a team of specialist agents.

It takes a list of messages: a system message that sets the role, then human messages from you, then AI messages from the model.

Most system prompts fail because they mix permanent instructions with facts that change weekly.

refusal_triggers = [ " ignore previous instructions " , " system prompt " , " reveal your prompt " , " bypass " , " jailbreak " , ] def check_input ( self , query : str ) -> Tuple [ bool , str ]: """ Validate user input.

" prompt = format_rag_prompt ( context , question ) assert context in prompt assert question in prompt assert " Answer based on the context " in prompt def test_agent_prompt_includes_tools (): tools = [{ " name " : " search " , " description " : " Search the web " }] prompt = format_agent_prompt ( tools ) assert " search " in prompt assert " Search the web " in prompt def test_no_prompt_injection

dumps ({ " equivalent_fps " : round ( 1000 / avg_step_ms , 1 ), " von_neumann_entropy " : round ( entropy , 4 ), " insight " : " Low entropy = converged on optimal path " if entropy < 1.0 else " High entropy = multiple solutions " }) The agent ( main.py ) uses Ollama's holo-9b-bf16 model with a soul.md system prompt that tells it when to call the VPU: # Core Identity You are a Hermes Agent with "h