epistemic-self-monitoring

Review-before-you-close self-check: a close with an edited file nobody reviewed in its medium (a document read, an image viewed, code run) is named, and the result is compared, point by point, with a source the agent did not write - the original, the specification, the real input, the owner's words. Shared skill + per-host hooks. Non-blocking by default, opt-in strict gate.

1 skill · hooks

Parts-ledger discipline: the parts of a request, the hardest first, each closed as done, blocked with an observed reason, or returned to the owner. Hooks count stubs and TODOs written per turn and flag deferred work with no closing ledger. Shared skill + per-host hooks. Never blocks.

1 skill · hooks

Plan-anchored drift self-check: shared skill + per-host cadence hook, installable as one unit.

1 skill · hooks

Structured-handoff discipline: the final message of a turn closes with a [HANDOFF] block modelled on SBAR / I-PASS - Status, Situation in the reader's terms, Options with a default, one Next action - so the reader can act without the trace. Hooks inject the format on the first closing-shaped tool call and flag a decision named but not handed off. Shared skill + per-host hooks. Non-blocking by default except a blocked close after a search that found nothing, which blocks once; opt-in strict gate.

1 skill · hooks

Change-scope self-check: the public functions a change reaches - an exported function, a helper other callers use, a sibling fixed for consistency - are recorded file by file, and the close names what the request does not and puts it back or to the owner. Shared skill + per-host hooks. Non-blocking by default except a block that contradicts itself, which blocks once; opt-in strict gate.

1 skill · hooks

Real-result discipline: when a service, a key or consistent tests are missing, the result is an error that says so, not a fallback that looks done. Hooks read each edit to product code for made-up data where a service was asked for (a rate table, a catch that answers, a missing key answered, a second host) or code that reads its caller, and each shell command for a change to the machine or a server started; each finding is said once, with what it means for the user. Shared skill + per-host hooks. No message on the prompt. Never blocks.

1 skill · hooks

Talk like caveman. Cut filler. Keep technical details explicit.

22 skills · 1 command · 3 agents · hooks

Persistent memory for Claude Code: the next session starts where the last one left off.

29 skills · 1 MCP server · hooks

tasteskill 2.5.1: professional company websites at the figma.com and x.ai level, with one section per screen, real photographs, real product UI, effects kept for tech and gaming pages, and a real-browser checker and finish gate.

1 skill

Open-source AI job search that runs locally in your AI coding CLI: scan job portals, evaluate listings into a structured A-H report with a 1-5 score, tailor your CV, track applications

5 skills · hooks

Cross-session memory and token savings for coding agents.

6 skills · 1 agent · 1 MCP server · hooks

Package any website or local web build into a lightweight desktop app with Pake CLI.

1 skill