Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/mishahanin/heading-os/implementnpx skills add mishahanin/heading-os --skill implementgit clone --depth 1 https://github.com/mishahanin/heading-osWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/mishahanin/heading-os/implement)<a href="https://agentmods.dev/skills/mishahanin/heading-os/implement"><img src="https://agentmods.dev/badge/skills/mishahanin/heading-os/implement.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00064 | $0.04406 |
| Opus 5 | $0.00032 | $0.02203 |
| Sonnet 5 | $0.00013 | $0.00881 |
| Haiku 4.5 | $0.00006 | $0.00441 |
Grade C, and why
implement scanned grade C with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured today.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Tells the agent never to refusehighAnti-refusal
Suppressing the ability to decline removes a core safety control; a later harmful request then succeeds.
This is a **soft** reminder, not a block. If Misha says proceed (or has already implied it), continue to Phase 0. Never refuse to implement on a MISSING result — CEO sovereignty holds. How it starts
The opening of the file, as written. The whole thing — 293 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Implement
Execute an implementation plan created by /create-plan. Read the plan thoroughly, execute each step in order, emit a structured trajectory for every phase boundary, and report on the completed work.
Version history (v1.3-v1.8): the emission discipline changed over several versions. Version 1.6 added typed flags (one Bash call per event, no temp file). Version 1.7 added the emit-time sequencing guard (exit 5) and the run-level --verify files reconciliation. Version 1.8 fixed --verify's blindness to bracketing defects (orphan wave_end reconciliation, out-of-bracket step_end detection, advisory flags for a wave_start missing its declared shape). The Trajectory emission contract below carries the authoritative invariants, restated in Phase 2, Phase 5, and the NEVER list. The full per-version changelog is in references/implement-details.md.
Trajectory emission contract
The trajectory (outputs/operations/implement/_trajectory_<run_id>.jsonl) is the input to /scrutinize trajectory:<run_id> audits (Agent-as-a-Judge, DevAI benchmark) and a verbatim audit record: emit through the helper only, never edit it. All emission is skipped under --no-trajectory. The invariants, enforced by the phases below and self-checked by --verify:
- Pairing. Every
step_starthas a matchingstep_end(same step number); step 0 (plan-load) is paired too. Everywave_starthas awave_end. - Sequencing (sequential waves / no-wave runs). Emit each
step_endbefore the nextstep_start; never open step N+1 while step N is open. The helper enforces this at emit time (v1.7). Astep_startopened while another step is still open exits5and does not land. Astep_endfor an unopened step does the same. To clear it, emit the missing marker first, or open a parallelwave_startfor legitimate interleaving. If work genuinely reorders, emit adeviationthat records the swap (the M1 discipline). Emit a step-scopeddeviationas well when a run defers a plan-declared terminal action. The final[0 0]push that a step's success criteria mandate is one such action. The same holds when a run reverses an approved decision mid-run, or diverges from explicit plan text at all. Even a trivial, functionally-identical substitution such as re-sourcing an import counts. Astep_endnote alone is not enough, because the structured trajectory must match the plan's Deviations section. Parallel-wave member steps may legitimately interleave inside their wave bracket, because an open parallel wave suspends the guard. - Wave bracketing.
wave_startprecedes the wave's firststep_start, andwave_endfollows all memberstep_ends.wave_end.successesequals the count of bracketedstep_endwith statusokordeviation, not the declared membership. Once a run emits any wave event, EVERY step but step 0 belongs to a bracket, and everywave_startcarries--step-countand--parallel/--no-parallel.--verifychecks all three since v1.8 (an orphanwave_endno longer buys itssuccessesclaim a free pass). - Literal paths.
--filetakes one literal path per real file touched: no globs, brace-shorthand, or count strings (the N1 discipline). - Whole-wave deferral. A wave not executed at all emits no
wave_start/wave_end; instead one wave-scopeddeviationkeyed to the wave's first step (--scope wave --wave N).
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- today Changed 93abf39fd321
- 4d ago First seen · 293 lines · 64 tokens per session scan C 2234c87005f6
implement is a skill published in the GitHub repository mishahanin/heading-os (11 stars, last pushed today), licensed Apache-2.0. It adds 64 tokens to every session and 4,406 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it C with 1 finding (tells the agent never to refuse). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
create-modal
Create declarative modals using the modal library API. Covers modal types (confirm, input, select, form), sections (Text, Buttons, Input, Textarea, Checkbox, List, Combo, When, Custom), rendering with OverlayModal, and keyboard/mouse handling. Use when adding modals or dialogs to the application.
merge-strategy
Git merge strategies, conflict resolution approaches, merge vs rebase recommendations, and branch integration patterns in sidecar. Covers pull strategy menu, direct merge workflow, squash merge, commit message templates, configurable defaults, and protected branches. Use when working on git merge features or making…
autoprompt
Explicit-only useful-first orchestration. Invoke /autoprompt to turn a mission into one executable roadmap, build dependency-safe lanes, and verify the result with independent reviewers. Never infer invocation from ordinary requests. Never resume from leftover artifacts without an explicit resume instruction.
autoprompt
Explicit-only useful-first orchestration. Invoke /autoprompt to turn a mission into one executable roadmap, build dependency-safe lanes, and verify the result with independent reviewers. Never infer invocation from ordinary requests. Never resume from leftover artifacts without an explicit resume instruction.
openlore-brainstorm
Transform a feature idea into an annotated story using a Domain Sketch or Constrained Option Tree. Use when asked to brainstorm, explore, or shape a feature before implementation.
ap-juror
L4 terminal leaf - G7 SIGN-OFF. One independent sign-off panel seat that saw none of the intermediate work. Binary PASS/FAIL on opened evidence; default-FAIL. A FAIL naming a P0/P1 blocker is NOT arbitrable into PASS.