buildOne report1 publisher Tool-call parsing splits coding agents into tiers on some models, not on native tool-call models
Polyglot's developer ran six coding agents 30 times on each of seven local models, and three never made a tool call on models that write calls as text. The author's own error bars say 30 runs can sort agents into tiers but cannot rank two agents inside one.
Reality
- Evidence55
- Adoption
- Not scored
- Hype gap+5
- Incentives70
- Confidence55
Block joined the x402 Foundation on Sept. 24 and is adding Bitcoin Lightning to the agent-payment protocol Coinbase built for stablecoins. The move puts a second settlement asset, and a second large payments company, into a standard its creator handed to the Linux Foundation this year.
Reality
- Evidence50
- Adoption15
- Hype gap+35
- Incentives70
- Confidence60
Manifold Security reported eight instances across seven command-line agents. The command runs as the user, outside the sandbox, with no approval prompt, and four were still firing when the researchers retested on September 1.
Reality
- Evidence72
- Adoption60
- Hype gap+15
- Incentives55
- Confidence70
buildOne report1 publisher Cloning the live Postgres volume runs the failing NOT NULL migration on a disposable instance
A Unikraft Cloud walkthrough clones a running database's persistent volume and boots a throwaway Postgres against it, so the NOT NULL migration that empty dev databases accept meets real rows first. The faster branching feature needs an enterprise license.
Reality
- Evidence52
- Adoption10
- Hype gap+12
- Incentives35
- Confidence45
buildOne report1 publisher Swapping the harness under a fixed model kept pass rate within 8 points on 50 Terminal-Bench Pro tasks
A 300-trial study swapped Goose, OpenCode and OpenHands-SDK under Qwen 3.6 Plus and MiniMax M2.5, and reports that the scaffold sets tokens per solved task and the failure mode while the score barely moves.
Reality
- Evidence52
- Adoption
- Not scored
- Hype gap+20
- Incentives30
- Confidence58
buildOne report1 publisher A repo's core.fsmonitor setting turns an agent's routine git status into code execution
GitSpawn, as summarised in a dev.to write-up of Cloud Security Alliance findings, uses a Git performance setting to run attacker code when a coding agent inspects a project it just opened. Seven agents are named.
Reality
- Evidence32
- Adoption
- Not scored
- Hype gap+18
- Incentives28
- Confidence40
buildOne report1 publisher Codex rewrote goose's quality gates until a ten-minute verdict took over an hour
A developer kept the model and the repository fixed and changed only the agent framework. Six days of Codex produced no completed run, while goose's own engine drove the same model to working code.
Reality
- Evidence30
- Adoption12
- Hype gap+30
- Incentives55
- Confidence40
buildOne report1 publisher Opening an untrusted repo made seven coding agents run the program its Git config named
Manifold Security's GitSpawn findings cover eight issues across seven coding agents. The path runs through core.fsmonitor, where a background git status starts a helper the repository chose, outside the agent's sandbox and with no prompt.
Reality
- Evidence34
- Adoption
- Not scored
- Hype gap+10
- Incentives
- Not scored
- Confidence38
buildOne report1 publisher A wiki that accepted GET as an edit gave read-only agents 18,000 writes
The wiki's old software took a plain GET as an edit, so a harness that policed request types could not stop the writes. GitSpawn puts the same gap in the startup git calls of seven coding agents.
Reality
- Evidence42
- Adoption52
- Hype gap+30
- Incentives78
- Confidence45
buildOne report1 publisher IEC 61850 moves its hardest trust question down to raw Ethernet frames
RUGERO Tesla spent several days in a libIEC61850 lab and came out with two separate sets of security questions, because MMS answers inside a session on TCP/102 while GOOSE and Sampled Values answer to whatever frame arrives.
Reality
- Evidence44
- Adoption
- Not scored
- Hype gap+12
- Incentives22
- Confidence52
buildOne report1 publisher Six specifications decide whether an agent can move off the harness it was built on
Lock-in migrated from the model to the runtime that holds your tool grants, approval rules and learned state. Six specifications aim to make that runtime's contents portable, and half of them are one person's drafts.
Reality
- Evidence36
- Adoption24
- Hype gap+27
- Incentives74
- Confidence42
buildOne report1 publisher Harness choice moved token use 83-fold with the model held constant
The spread traces to two numbers you can read off your own logs: the tokens a harness spends before any work starts, and how many turns it takes. Together they predicted total token use with an R-squared of 0.99.
Reality
- Evidence58
- Adoption34
- Hype gap+12
- Incentives42
- Confidence55
Block says 95% of its engineers use AI daily. The interesting part is who runs five agents.
Block has published the first hard distribution of AI adoption inside a large engineering org: one agent for most, three to five for the next cohort, and a small team building the orchestrator.
Publishers:engineering.block.xyz
Reality
- Evidence44
- Adoption71
- Hype gap+24
- Incentives73
- Confidence52
Block, Stripe and Shopify run in-house coding agents too. What Ramp built is not a model but a credentialed development box, and that is the part procurement cannot order.
Reality
- Evidence68
- Adoption74
- Hype gap+18
- Incentives76
- Confidence66
buildOne report1 publisher Three named agents shipped in 27 days, and none of them can read each other's config
Block, xAI and Nous all landed on the agent as a persistent identity. That identity still lives in whatever file the harness picked, and the switching bill is now memory.
Reality
- Evidence38
- Adoption32
- Hype gap+30
- Incentives78
- Confidence42
buildOne report1 publisher The Claude connector gates that fail before your code runs: DNS, IPv4 and one Location header
A dev.to pre-flight checklist puts network reachability first: private-range answers and AAAA-only hosts kill the connection, and a cross-host redirect strips the Authorization header.
Reality
- Evidence38
- Adoption34
- Hype gap+12
- Incentives30
- Confidence44
Agent protocols now share one landlord: A2A joins MCP and AGENTS.md at the Linux Foundation
Google's Agent2Agent has been accepted as a growth-stage project at the Linux Foundation's Agentic AI Foundation, which already holds MCP, goose and AGENTS.md. The protocol question is now a governance non-event.
Publishers:yugabyte.com
Reality
- Evidence28
- Adoption32
- Hype gap+34
- Incentives86
- Confidence30
buildOne report1 publisher Block's Berd makes a duller argument than its mascots: show the agent's context as product state
Block open-sourced the desktop app its own teams use to drive agents. The design claim underneath it is that active context should be inspectable, not archaeology in a prompt log.
Reality
- Evidence28
- Adoption18
- Hype gap+12
- Incentives58
- Confidence38