Posts
Independent essays on AI-native work patterns, agent infrastructure, and what actually works when running agents as real tools.
-
1Password op run Makes Vault-Backed Credentials Tool-Transparent
Updated:At process launch, op run resolves 1Password references into ordinary environment variables, so CLIs and MCP servers need no vault-specific code.
-
A Never-Say List Is the Fingerprint of an AI Voice
Updated:Writing samples teach cadence. Corrections and mode-specific never-say lists encode the rejected choices that make an AI voice distinctive.
-
Docs-as-Code as a Knowledge Distribution Stack
Updated:Markdown, Git, llms.txt, and MCP create human and agent read paths from one source, while AI-mediated pull requests remove Git fluency from writing.
-
Encode Workflow Failure Modes, Not Only Successes
Updated:A leaked folder name exposed a safety-check gap. Encoding that failure class in the workflow makes it visible in every downstream run.
-
The MCP Proxy, Live: 206 Tools in 2,100 Tokens
Updated:A live MCP proxy reduced startup context from about 217,000 to 2,100 tokens across 206 tools. Most sessions now run from a ten-tool always-on set.
-
When AI Makes Proof Cheaper Than Persuasion
Updated:AI makes working evidence cheap enough to replace part of organizational persuasion: build the prototype, let people test it, then decide.
-
The Right LLM for Each Step in an Agent Pipeline
Updated:Match model tier to cognitive load. A fast, single-call path with pre-fetched context cuts a 25-30 second two-call response to under 10 seconds.
-
Git-Native Agent Skills Need a Release Lifecycle
Updated:Tessl adds versioned publishing, immutable installs, and deliberate updates to repo-based skills. CI review exists, but teams must require it.