Skip to content
Skills Library
AI & agents · Review & audit

instructions-method

/instructions-method

Create, audit and maintain durable project instructions in AGENTS.md, CLAUDE.md or the target host's instruction file. TRIGGER WHEN: auditing, creating or updating AGENTS.md, CLAUDE.md or durable project instruction files.

Licence
MIT
Last repository change
27.10.2026

Assessment and security

95/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content30 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
7
Scripts
0
HTML comments
2
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 27.10.2026, without references to retired models or APIs.

Licence

MIT, open: use and modify while retaining author attribution.

What to check before installing
  • SKILL.md zawiera komentarze HTML (2). Nie widać ich w podglądzie Markdown, a model je czyta.

What it does and when to use it

Create, audit and maintain durable project instructions in AGENTS.md, CLAUDE.md or the target host's instruction file. TRIGGER WHEN: auditing, creating or updating AGENTS.md, CLAUDE.md or durable project instruction files.

Original author description

Create, audit and maintain durable project instructions in AGENTS.md, CLAUDE.md or the target host's instruction file. TRIGGER WHEN: auditing, creating or updating AGENTS.md, CLAUDE.md or durable project instruction files.

How it works

  1. Destination and authority
  2. Preparation
  3. Audit and create
  4. Apply and verify
  5. Scientific knowledge pointer
  6. Loop block
  7. Coordinator delivery
Original SKILL.md headings · 182 lines, 1,541 words
SKILL.md preview
---
name: instructions-method
description: >
  Create, audit and maintain durable project instructions in AGENTS.md, CLAUDE.md or the target host's instruction file.
  TRIGGER WHEN: auditing, creating or updating AGENTS.md, CLAUDE.md or durable project instruction files.
---

<daodan_host_binding host="claude">
This binding governs execution of this package and its loaded references. It takes
precedence over executable examples written for another host. Use only registered
tools and their actual input schema; an example never creates a tool, argument,
permission, model alias or nested worker capability. Preserve the generated worker
allowlist, dependency closure, scope and delivery contracts.

Claude binding: repository tools are Read/Glob/Grep and Write/Edit; shell execution
uses Bash and network fetch uses WebFetch. Pass the executable string in Bash's
command field, and use its registered timeout/background schema. Worker dispatch
uses the available native Agent mechanism and its registered agent-selection and
prompt fields. Do not turn a prose role label into an invented role argument.
Inherit the session model unless the declared workflow and active schema permit an
override. Load installed skills by qualified ID; named roles use their owning
plugin's agent identity, while workflow entries use /<plugin>:<workflow>.
Required foreign providers use Claude's marketplace/install commands:
claude plugin marketplace add <source>, then claude plugin install <provider>@<marketplace>.
A provider that is installed but disabled is enabled with
claude plugin enable <provider>@<marketplace>; check `claude plugin list --json`
before the first phase, and stop with that command when any plugin of this
package's dependency closure, direct or transitive, is off.
A declared package MCP server is configured through the generated MCP manifest;
still check its tools are actually available in the active session.

The executing agent and its coordinator own continuation, recovery, verification
and delivery. Skills supply knowledge and method instructions; they are not
executors or a runtime persistence mechanism.

Resolve each component by kind and exact installed identity. A role is dispatched
as a worker, a skill loads instructions, and a workflow starts its declared method.
Kernel owner-qualified skill references are logical provider identities: resolve
them to the actual host registry name and verified installed provider path. Do not
invent a registry string by adding an owner's prefix when the host does not use it.
Do not load an agent identity as a skill or execute a workflow name as a shell
command. Declared required foreign providers remain mandatory. If a required
provider or capability (including required MCP access or isolation) is unavailable,
stop its dependent phase and report the missing capability. Optional external
providers follow the method's declared selection and absence behavior; their
absence alone does not block the base workflow. Report any explicitly requested
but unavailable provider contribution as unfulfilled and continue independent
work. Do not install a different host's plugin or
substitute unrelated skills to claim coverage.

Shell execution alone does not establish interactive stdin. Before a capture or
checkpoint, use an available session-control mechanism. When stdin is unavailable,
keep user-driven capture in a supported background process with bounded output and
an explicit stop signal controlled by the coordinator after the user's completion
message. Close owned browser/process resources on completion or cancellation.

Continue working within the current authorization while the requested result is
achievable. After a tool error, inspect its cause, correct the relevant input or
implementation, retry when justified and verify the outcome; never repeat a
failing action blindly. Do not close the turn with a promise or a partial result
when an authorized next step can still produce or verify the requested result.
Stop only for a documented real impediment, a missing required capability, an
unresolved human decision or the caller's stop request. At that boundary, report
completed results, remaining work and the concrete prerequisite for continuation;
continue any independent work that the blocker does not prevent.

A tool's successful exit, a worker stop or an idle session is not the deliverable.
Return final text naming observed findings, artifact paths, executed checks and
blockers. An empty worker result is missing delivery: inspect or resume that worker
when permitted, or record failure. Verify required artifacts and candidate gates
before marking the requested outcome complete.

This is an execution binding, not a rewrite of source examples. Preserve programming
language fences, domain code and host APIs that are the subject of marketplace-ops
or agent-sdk-builder. A host-as-subject API is not a tool provided to this session.
Compiled binding and static checks do not prove an installed host executed a run.
</daodan_host_binding>

# Durable instructions method

Load the skill `work-controller:work-record` for run records and gates.
Receive project root, worktree/snapshot, scope, requested instruction destinations,
mode, existing files, evidence paths and authorizations. Reuse the caller's run.

## Destination and authority

Read `references/instruction-loading.md` during preparation. Bind the actual
host/version, configured discovery and applicable root/nested scopes; a file's
presence does not prove it is loaded or enforceable. Record evidence and unknowns
in the owned run rather than assuming every host reads the same files.

Use explicit user destinations first, then the project's existing authoritative
instruction file and documented synchronization rule. For a new project, use the
mechanism confirmed for the active host/configuration (for example AGENTS.md for
Codex, CLAUDE.md for Claude, or configured rules). If the binding is unknown, present concrete
destinations before writing; do not invent a universal filename. Multiple hosts
may use generated or thin-pointer copies only under an explicit project rule.
Never maintain two independent contradictory instruction sets. Honor nested
scopes and do not rewrite a generated instruction copy directly.

## Preparation

Read manifests, source structure, entry points, configured commands and CI before
checking claims. Map evergreen categories and ownership rather than every file.
Read approved requirements/decisions separately from implementation. A supplied
Codebase Map is consumed through `codebase-map:map-method`, bound to its exact run and
snapshot, preserving unknown/disputed claims and inherited premise provenance.
For test guidance, derive the policy from this project's configured commands,
services, conventions and approved behavior. Check its completeness with
`references/test-suite-rules.md`; do not turn a template into project authority.

## Audit and create

Dispatch `instructions-auditor` with explicit audit-only or create mode and the
inputs above. Audit-only returns findings to the owned run, without editing any
durable file. Use its claim/reference/duplication checks; do not add another
detector prompt to the caller. Creation uses a planned destination and verified
facts, not invented project policy.

Read `references/working-principles.md` when a new project needs working guidance
and `references/test-suite-rules.md` when tests exist or are planned. Existing
equivalent rules remain valid; missing canonical wording alone is not a defect.
Read `references/output-economy.md` when creating or maintaining context/output
guidance or instructions for projects that author agent components or tools.
Keep the always-loaded rule short and preserve required reading,
independent verification and domain report contracts.
`references/instruction-example.md` illustrates structure, not current package
versions or requirements for the target project.

## Apply and verify

Audit is the default. With authorized fix/commit, reuse the same claim report and
read `${CLAUDE_PLUGIN_ROOT}/skills/project-knowledge/references/apply-method.md`.
Preserve unverified claims without inventing
support; distinguish them from disproved claims. Resolve uncertain intent before
changing an obligation. Extract repeated detail to existing docs with a working
pointer when authorized; keep the primary entry point coherent. Recheck paths,
commands, instruction scope and generated-copy parity before recording completion.
Apply the project's synchronization rule from its canonical owner. Distinguish
verified generated parity from instruction activation: use a supported reload or
new session when required, or record the unperformed loading check and its effect
on the active run. Never claim a textual obligation is mechanically enforced
without evidence of the actual guard in the relevant environment.
Run status, task lists, temporary paths and one-run results stay in the operational
record, never in durable instructions.

## Scientific knowledge pointer

When a project adopts scientific research and instruction changes are authorized,
keep a concise durable pointer to `docs/research/README.md` and
`research:science-based-approach` in its canonical instruction file. Explain that
mode is read at runtime, relevant evidence is revision-qualified and limits stay
visible. Require existing related Markdown to be searched before new research;
returning topics update the same file with timestamps, visible supersession and
descriptive changelog, retaining history separately. Preserve unrelated
instructions and apply the project's synchronization
rule. Never persist the current ON/OFF state, check date or run progress here.

## Loop block

A project driven by the lifecycle loop carries one bounded block in its
authoritative instruction file, between `<!-- LOOP-HARNESS:BEGIN -->` and
`<!-- LOOP-HARNESS:END -->`. The block holds only what every task needs at
its start: the verified test, lint, type and build commands with their
prerequisites; on a project that keeps decision records, the active ones per
domain, by identity and title, supplied by the caller as resolved active decision inputs; the
workspace convention (worktree root, branch naming); the neutral specification
destination (`docs/specs/` unless explicitly configured otherwise); and routing:
design/brainstorm requests use Daodan's delivery method, which writes and obtains approval of
the specification before planning, reusing existing answers. This is an instruction,
not proof of automatic host activation. State that run records live under
`.daodan/runs/`. Reasons stay in the records,
procedures stay in the loaded methods, and run state never enters the block.

Create the block at project initialization when creation is authorized;
otherwise report it as a planned change. At the close of a task, regenerate
the block from the run record and, when the project keeps decision records, the
record index, and touch nothing outside
the markers. A missing marker pair in an existing file is a finding, not a
reason to append a second block. Generated copies follow the project's
synchronization rule as for any other instruction change.

## Coordinator delivery

For requested HTML or useful comparison/navigation, the coordinator reads
`${CLAUDE_PLUGIN_ROOT}/skills/project-knowledge/references/human-view.md` before
closing the run. Instructions-auditor deliveries stay native.
Files in this package (7)
  • SKILL.md 11 KB
  • SKILL.toml 1 KB
  • references/instruction-example.md 24 KB
  • references/instruction-loading.md 7 KB
  • references/output-economy.md 7 KB
  • references/test-suite-rules.md 9 KB
  • references/working-principles.md 9 KB

Versions

RepositoryScoreLicence
This page · acaprino/alfio-claude-plugins95MIT
acaprino/alfio-claude-plugins95MIT
acaprino/alfio-claude-plugins95MIT
acaprino/alfio-claude-plugins95MIT
acaprino/alfio-claude-plugins95MIT
acaprino/alfio-claude-plugins95MIT

Identical content appears in 1 other GitHub folders. We show one copy with the highest score.

Database snapshot: 13.11.2026 · package version d0f4e0d36df2 (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 100agents-md
    AGENTS.md: create, audit, or maintain the file. Use when the user mentions AGENTS.md, project instructions, instruction budget, or progressive disclosure of agent instructions.
  • 100agents-md-author
    Bootstrap or update project AI-agent instruction files (AGENTS.md for Codex, CLAUDE.md for Claude Code, or both) from the current project's structure. Triggers on: write AGENTS.md, write CLAUDE.md, generate AGENTS.md, agent instructions, project conventions for AI, bootstrap…
  • 100target-instructions-bootstrap
    Create the target repository's local AI instruction scaffold during first-use bootstrap. Use from repo-init-scan when `.github/instructions`, runtime adapters such as `AGENTS.md` / `CLAUDE.md`, or the neutral project facts scaffold are missing. Do not use to overwrite existing…
  • 90writing-project-instructions
    Use when authoring, editing, updating, trimming, or adding rules to a CLAUDE.md, AGENTS.md, CLAUDE.local.md, or other project-level instruction file for an AI coding agent — including small one-line changes. Triggers on "write a CLAUDE.md", "audit my CLAUDE.md", "trim my…
  • 100agents-md-authoring
    Create and maintain AGENTS.md files that bootstrap AI agent context. Use when a project lacks an AGENTS.md, when onboarding a repo for AI agents, or when updating project conventions agents should follow. Covers the agents.md standard, the context-loading instruction block, and…
  • 100init-agent-config
    Use if creating, auditing, or migrating CLAUDE.md/AGENTS.md/REVIEW.md instruction files.

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.