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/evoelsewhere/evoflux/easd-implementnpx skills add evoelsewhere/evoflux --skill easd-implementgit clone --depth 1 https://github.com/evoelsewhere/evofluxWhat 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 | $0.00051 | $0.00861 |
| Opus 5 | $0.00026 | $0.00430 |
| Sonnet 5 | $0.00010 | $0.00172 |
| Haiku 4.5 | $0.00005 | $0.00086 |
Grade A, and why
easd-implement scanned grade A with 0 findings 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 2d ago.
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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 74 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Implement an active EASD run
Repository contract
Read .evoflux/easd/config.json and its rules_file before phase work. Treat
the injected EASD context as the authoritative current Run; when the owning
runtime is accessible, corroborate it under runtime_directory. An isolated
worktree intentionally has no checkout-local runtime copy. Resolve the accepted
Spec through the common specs/ index and confirm it matches the Run snapshot. Use features/,
architecture/, and reference/ only for living knowledge the accepted change
actually affects; existing external docs remain in place. Repository documents
are the shared source of truth. Stop on a missing tracked contract, stale hash,
generation conflict, or unavailable injected runtime; never reconstruct
authority from chat memory or a stale database-only projection. For planned,
bind work to the accepted Plan mission. For direct, bind only to accepted
Spec/AC/scope and never invent a Plan identity.
Implement only the active accepted contract and the mission or ownership scope assigned to this agent. EvoFlux runtime checks remain authoritative even when this skill is active.
State gate
Before each implementation or resumed turn, re-read the injected run ID,
active status, accepted Spec hash, delivery flow, and mission contract. For
planned flow, also re-read the accepted Plan hash/mission and stop if either is
absent or stale. For direct flow, stop if any Plan identity is claimed. Stop if
the mission was cancelled/reworked or durable context is unavailable. Record
the dirty-worktree baseline so user/peer changes are not claimed as this output.
Execution discipline
- Confirm the injected run ID, spec hash, owned ACs, target repositories and paths, dependencies, constraints, and evidence policy before mutation.
- Read the nearest applicable
AGENTS.mdand existing implementation/tests. Preserve unrelated and concurrent work in the shared repository. - Make the smallest coherent change that satisfies the owned ACs. Keep public contracts, persistence, security, and compatibility aligned across affected layers.
- For an observable behavior change or bug, add the focused regression first when practical and confirm it fails for the expected missing behavior. Make the minimal implementation, confirm the regression passes, then refactor only while it stays green. When test-first is inapplicable, record the alternate baseline/evidence instead of pretending a red-green cycle occurred.
- Run the focused accepted verification commands that are safe and available,
plus boundary-specific regression checks required by repository instructions.
Prefer canonical commands from
AGENTS.md, project manifests/task runners, CI, and existing test configuration over invented shell pipelines. Read the fresh output and exit status before making any pass claim; an unavailable or unknown required command is an evidence gap, not an implicit pass. - Report changed paths, checks and exact outcomes, evidence gaps, and each owned AC result. A task checkbox, handoff, or agent confidence is progress—not trusted evidence—and must not be presented as convergence.
What ships with it
1 file 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.
- 2d ago First seen · 74 lines · 51 tokens per session scan A 24e15cf3f1ba
easd-implement is a skill published in the GitHub repository evoelsewhere/evoflux (5 stars, last pushed 6d ago), licensed Apache-2.0. It adds 51 tokens to every session and 861 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
skill-creator
Scaffold, author, and validate an Agent Skill. Use when the user wants to create a new skill, turn a repeatable workflow into a reusable SKILL.md, or fix a skill that fails to load.
rove
Use when controlling Rove tasks, parallel coding attempts, hosted agent sessions, task lifecycle, or the daemon-owned issue tracker from a shell. Also the ONLY channel for messaging another agent session on this machine — rove api send, never a peer/MCP side channel.
openai-docs
Use when the user asks how to build with OpenAI products or APIs, asks about Codex itself or choosing Codex surfaces, needs up-to-date official documentation with citations, help choosing the latest model for a use case, latest/current/default-model prompting guidance, or model upgrade and prompt-upgrade guidance; use…
skill-creator
Create or update AnkaLoop skills. Use when designing, structuring, or packaging skills with scripts, references, and assets. This skill should be used when users want to create a new skill (or update an existing skill) that extends AnkaLoop's capabilities with specialized knowledge, workflows, or tool integrations.
telegram-sender
Send and edit Telegram messages, and send photos via Bot API. Use when AnkaLoop needs to send text, reply to a specific message, edit an existing message, send an image, or push proactive notifications (cron results, heartbeat alerts, task status). Requires ANKATELEGRAMBOTTOKEN env var.
Nori Skillsets
Use when the user asks about anything related to nori.