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/rakovi4/continue-framework/continuenpx skills add rakovi4/continue-framework --skill continuegit clone --depth 1 https://github.com/rakovi4/continue-frameworkWhat 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 | $0.00029 | $0.05552 |
| Opus 5 | $0.00015 | $0.02776 |
| Sonnet 5 | $0.00006 | $0.01110 |
| Haiku 4.5 | $0.00003 | $0.00555 |
Grade A, and why
continue 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 — 170 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/continue - Resume Development
Workflow
- Identify work item from argument
- Backlog promotion -- if the story row is in the Backlog table in
ProductSpecification/stories.md, move it to In Progress before proceeding - Read progress file, bootstrap if missing (stories only —
.claude/templates/workflow/bootstrapping.md). Missing means both locations came back empty — resolveProductSpecification/stories/NN-story-name/andProductSpecification/stories/done/NN-story-name/before the bootstrap may fire (.claude/rules/workflow.md, "Resolving a story folder"); a hit in either is the story, so read itsprogress.mdand bootstrap nothing. What an unchecked bootstrap does to an already-shipped story is in the template's "Precondition: the story genuinely has no folder" For an existing task, normalize legacyType: bugtobugfixandType: refactoringtorefactorin memory; do not rewrite its type or folder merely to migrate terminology. - Find next step -- run the plan-integrity check (
.claude/templates/workflow/plan-integrity-check.md) overprogress.mdfirst; on a failed check report it and STOP without dispatching. Otherwise the next step is the first[~]or[ ]entry. A task with none runs its unconsumed terminal review, or archives immediately when its work log already recordsTask review: consumed. - Read relevant context -- read
carryover.md(if present), the current scenario's summary, and onlyworklog/records matching the current heading/step or named by its compact pointer. Do not load the directory wholesale./continuenever writes journey files (/handoffis their sole writer). - Load ADR context -- check for
decisions/*-decision.mdfiles in the story directory. If any exist AND the current step references the ADR (via "see ADR" annotation or matching scenario), read it. ADRs contain architectural decisions, schema changes, edge cases, and implementation guidance that the work unit needs. - Execute one work unit -- dispatch sub-skills per tables below. If no named row matches, execute the checkbox intent directly in the main agent.
- Discovery gates -- when the next step is
[ ] adapters-discovery, map usecase ports to adapters (see.claude/guidelines/workflow-detail.md), then insert concrete steps. A TDD task'ssteps discoveryinserts its scoped TDD plan; a refactor task'srefactor (steps discovery)reconciles direct## Workcheckboxes, not merely a companion artifact. - Update records -- create one
worklog/Markdown record for this invocation (maximum 200 lines) and keep routine evidence, notes, coordinator plans, checkpoints, and proposals there. Composite stages update this active record while in flight. Keepprogress.mdto headings, permitted compact tokens, and one-line checkboxes of at most 200 characters: mark completed and advance next. Add a work-log pointer only when a later decision must consume that exact record. Any new checkbox must pass.claude/templates/workflow/finding-admission-test.md; a failing finding is reported and stored nowhere. - Update stories.md -- for stories only, update the phase columns in
ProductSpecification/stories.md(see below) - Behavior commit -- include
progress.md, the invocation's work-log record, andProductSpecification/stories.mdfor stories. Before every progress commit, stage it and run plan-integrity check 8 against its zero-context staged diff; only added lines are checked, so legacy prose is grandfathered. Re-home any rejected evidence in the active work-log record and re-stage. For completion, also confirm the staged advance and zero[ ]/[~]entries. Stories archive in that commit; tasks remain active until terminal review. - Refactor and review -- land any owed
/refactorseparately. Stories review at block boundaries. Tasks skip per-step review and, once all planned work joins, dispatchagent-review-agentandpremortem-agentconcurrently over the whole task history before archive. - Triage & auto-fix -- partition findings per
triage-and-auto-fix.md, resolve any permitted quiz, and append every verdict/disposition to the active work log. A task record also addsTask review: consumed; findings never trigger a second terminal batch. Land SAFE fixes inreview-fix:and records without fixes inworklog:.
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 · 170 lines · 29 tokens per session scan A a5fd7e98012e
continue is a skill published in the GitHub repository rakovi4/continue-framework (50 stars, last pushed 5d ago), licensed MIT. It adds 29 tokens to every session and 5,552 once invoked, about $0.0001 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 2026-08-30.
Other skills, from other repositories
wayfinder
Plan a huge chunk of work (more than one agent session can hold) as a shared map of decision tickets on your issue tracker, and resolve them one at a time until the way to the destination is clear.
setup-matt-pocock-skills
Configure this repo for the engineering skills: set up its issue tracker, triage label vocabulary, and domain doc layout. Run once before first use of the other engineering skills.
accessibility
Consolidated accessibility skill entrypoint for WCAG 2.2, ARIA Authoring Practices, cognitive accessibility, Section 508, EN 301 549, design intent verification, and the Accessibility Planner workflow.
release-notes
Generate user-facing release notes from tickets, PRDs, or changelogs. Creates clear, engaging summaries organized by category (new features, improvements, fixes). Use when writing release notes, creating changelogs, announcing product updates, or summarizing what shipped.
retro
Facilitate a structured sprint retrospective — what went well, what didn't, and prioritized action items with owners and deadlines. Use when running a retrospective, reflecting on a sprint, creating action items from team feedback, or learning how to run effective retros.
bug-triage
Read all open bugs in production/qa/bugs/, re-evaluate priority vs. severity, assign to sprints, surface systemic trends, and produce a triage report. Run at sprint start or when the bug count grows enough to need re-prioritization.