@timiagama/ecr 0.0.1 → 0.1.0

package/LICENSE

@@ -0,0 +1,202 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright 2026 Timi Agama
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

package/NOTICE

@@ -0,0 +1,16 @@
+Explicit Constraint Referencing (ECR)
+Copyright 2026 Timi Agama
+
+This product includes software developed by Timi Agama.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.

package/README.md

@@ -1,11 +1,415 @@
-# @timiagama/ecr
+# ECR — Explicit Constraint Referencing
 
-This version reserves the name. There is nothing to install yet.
+**A Markdown convention that helps reduce coding agent errors by turning
+cross-references in architecture docs into a graph that coding agents navigate 
+with `grep`.**
 
-**Explicit Constraint Referencing (ECR)** is a Markdown convention, and a linter
-for it, that turns the cross-references in architecture documents into a graph a
-coding agent can navigate with `grep` alone — no index, no server, no plugin.
+Agents gain IDE-like "Go to Definition" and "Find all references" across large
+document corpora, reducing missed constraints and context-window bloat.
 
-The first working release is `0.1.0`, published from CI with provenance.
+No server. No index. No embeddings. Just numbered headings and a
+`## References` section.
 
-Until then, see [github.com/timiagama/ecr](https://github.com/timiagama/ecr).
+---
+
+## Why coding agents miss constraints
+
+Your architecture documentation contains constraints. Some govern other
+constraints. Some impose limits. Some define contracts that other components
+must satisfy.
+
+You understand those relationships when you read the docs. A coding agent does
+not — because ordinary Markdown does not encode them in a form an agent can
+deterministically traverse. They live in your head, in the ordering of files,
+in the fact that you happen to remember that the retry policy in one document
+constrains the payment flow in another.
+
+Without stable identifiers and explicit references, the agent either pulls
+more documentation into context hoping the relevant constraint is there, or
+proceeds with what it has found and silently violates a constraint written
+somewhere it never looked.
+
+The second failure is the expensive one, and it is easy to miss in local
+review. Nothing in the file you are reading points at the thing you missed.
+
+## ECR exploits how coding agents already work
+
+A coding agent does not read your whole repository into its context window.
+
+Its harness gives it tools to inspect the repository on demand: list files,
+search for text or symbols, open the relevant files, read specific sections,
+and run commands.
+
+When an agent works on code, it has stable identifiers: function names, type
+names, imports, filenames, symbols. It can search for `retryWithBackoff`, find
+where it is defined and used, read a handful of relevant files, and never load
+the repository into context.
+
+That is how an agent can operate on a codebase far larger than its context
+window. The model builds the context it needs as it works.
+
+Code already gives the agent stable identifiers to search for. Architecture
+prose usually does not.
+
+The "retry policy" is "retry semantics" in one document, "backoff rules" in
+another, "the orchestration contract" in a third. There is nothing exact to
+search for.
+
+ECR gives architecture documentation the same kind of searchable structure
+that code already has. Stable identifiers and explicit references turn "what
+constrains this?" into something the agent can actually look up.
+
+## What ECR does
+
+Your IDE gives you two operations on code: **Go to Definition** and **Find All
+References**. ECR gives your architecture documents the same two operations,
+with `grep` doing the job the IDE's index does, and with nothing to build or
+keep in sync.
+
+| IDE operation | ECR equivalent | Answers |
+|---|---|---|
+| Go to Definition | forward reference | What does this document rely on or answer to? |
+| Find All References | reverse reference | What elsewhere relies on or refers to this? |
+
+> General-purpose links help you navigate from a known source to a known
+> destination. ECR adds something different: stable section identities and
+> explicit relationships that an agent can traverse mechanically in both
+> directions.
+> 
+> That means the agent can **retrieve a constraint from the specific section
+> that defines it rather than loading the document that contains it**. And it
+> can search backwards to discover everything elsewhere in the corpus that
+> refers to that constraint.
+
+The two directions solve different problems. Forward references surface the
+constraints a document knows it must obey. Reverse references surface
+everything elsewhere in the corpus that relies on it — relationships the
+target has no reason to know exist, and each one a limit on what you can
+safely change.
+
+In practice, both operations are one search.
+
+For example, one document sets a constraint. The sub-heading containing the constraint has
+a stable identifier `8.1#3.2`:
+
+```markdown
+### 8.1#3.2 - Bounds
+
+At most five attempts. A stage that exhausts its attempts fails permanently
+and is reported, not re-queued.
+```
+
+Another document, elsewhere in the corpus, depends on that constraint and
+cites it by the same stable identifier:
+
+```markdown
+A discrepancy halts settlement for that merchant only. Orchestration-level
+retry does not apply — see 8.1#3.2.
+```
+
+They need not share a folder, a filename convention, an index or a database.
+The stable identifier is enough to make the section-level link searchable, so
+finding either end from the other is one search:
+
+```bash
+grep -r "8.1#3.2" docs/
+```
+
+That is the mechanism. The identifier is exact, so the search is exact: it
+returns the section that defines the constraint and every document citing it,
+wherever they sit. Nothing is indexed, nothing is embedded, and there is no
+ranking to be wrong about.
+
+Nobody loads a repository into their head to use **Go to Definition**. The
+index does the work and you read only what it returns. ECR works the same way:
+the edges are plain text on disk, so an agent greps and pulls the few sections
+that matter instead of the whole corpus. **The graph lives in your filesystem,
+not in your context window.**
+
+## The four rules
+
+**1. Every document has a numbered H1.** That number is its DocID — its
+identity, independent of title or path.
+
+```markdown
+# 4.2 - Payment Processing Contract
+```
+
+**2. Every sub-heading number begins with the DocID.**
+
+```markdown
+# 4.2 - Payment Processing Contract
+## 4.2#1 - Idempotency
+### 4.2#1.1 - Key Derivation
+```
+
+A sub-heading number is a SectionID, and it is self-locating: `4.2#1.1` is
+document `4.2`, section 1, sub-section 1. The `#` marks where the document
+identifier ends and the section path begins, so `4.2#1` and `4.2.1#1` are
+different things and neither can be mistaken for the other.
+
+**3. Inline references use `see` or `per`.**
+
+```markdown
+Retry semantics are applied per 8.1#3.
+```
+
+Not "as described in the orchestration doc." A number, so it can be found.
+
+**4. Every document declares its external references** in a `## References`
+section, with a typed direction and a reason.
+
+```markdown
+## References
+
+- 3.1 - Prompt Contracts (authority - defines the guardrail criteria enforced here)
+- 8.1 - Orchestration Contract (constraint - retry semantics applied to evaluation runs)
+```
+
+That's the whole convention. Four rules, no new syntax, no tooling required to
+adopt it.
+
+### Direction semantics
+
+| Direction | Meaning (ECR) | Agent action (recommended) |
+|---|---|---|
+| `authority` | the cited document governs this one | Read it first. Treat it as the governing source. Flag apparent conflicts. |
+| `constraint` | the cited document restricts this one | Check your change doesn't violate it. |
+| `contract` | the cited document defines interfaces or obligations | Conform exactly. |
+| `dependency` | this document consumes the cited one | Read it to use it correctly. |
+
+### Searching precisely
+
+A literal search shows the mechanism. Two refinements make it reliable:
+
+```bash
+# Go to Definition — locate the section a reference names
+grep -rInE "^#+ 8\.1#3([^0-9]|$)" docs/
+
+# Find All References — every document citing that section or its sub-sections
+grep -rInE "([Ss]ee|[Pp]er) 8\.1#3([^0-9]|$)" docs/
+```
+
+The reverse search is where ECR exposes relationships the target section
+cannot know about. On the [example corpus](examples/docs/) it returns five documents across
+four folders, three of which cite `8.1#3.2` specifically: downstream
+documents a change to the retry bounds could break. One of them, settlement, appears nowhere in the orchestration contract.
+
+Match `[Ss]ee` and `[Pp]er`, not just lowercase: a reference at the start of a
+sentence is still a reference, and a pattern that misses it under-reports the
+exact thing you are searching for. The `([^0-9]|$)` ending stops `8.1#3` also
+matching `8.1#30`. Matching a whole document rather than its sections needs a
+stronger guard, because `.` and `#` are not word boundaries — the supplied
+[coding agent navigation protocol](protocol/navigation-protocol.md) carries the tested pattern.
+
+## Why `grep` is enough
+
+What ECR produces is a directed graph: documents and sections are nodes,
+cross-references are edges, and document-level edges carry relationship types.
+
+The useful property of a graph is that you can answer questions about a node
+by following its edges, without reading every other node. That's the property
+being borrowed. `grep` may scan every file, but the agent only reads what the
+edges lead it to. An agent asking "what governs this component?" doesn't read
+your entire documentation corpus — it resolves an identifier and follows
+edges, and every edge it follows is one it can justify rather than infer.
+
+It matters that the identifiers are stable and the edges are plain text. Exact
+identifiers make retrieval deterministic: against the same corpus, the same
+query returns the same sections every time, with no ranking, no similarity
+threshold, and no chance of a plausible-looking near-miss. Plain text on disk
+means the traversal tool is `grep`, which every coding agent already has.
+
+## Evidence from one project
+
+Here is my experience with it: **one project, my own assessment, not a benchmark**.
+
+I converted a documentation corpus to ECR: **47 Markdown files**, 40 carrying
+`## References`, **1,672 numbered headings**, and **575 inter-document references** —
+413 References entries plus 162 inline `see`/`per` citations, 95 of them
+section-precise.
+
+Then I pointed a coding agent at the [navigation protocol](protocol/navigation-protocol.md), told it to use ECR,
+and started building. It produced around **19,000 lines of TypeScript across
+124 source files** with **34 test files**. Retrieval was `grep`. There was no
+server, no index, no embedding model and no MCP tool in the loop — the agent
+read the protocol, walked the references, and pulled the sections it needed.
+
+In my judgement the result contained unusually few errors for a build of that
+complexity. That's a subjective claim about an unpublished, commercially
+confidential corpus, and you should weight it accordingly. The mechanism
+underneath it, though, is not subjective — you can check that on your own
+documentation in about two minutes.
+
+## Getting started
+
+### Commands
+
+```bash
+npx @timiagama/ecr init ./ecr     # install ECR documentation
+npx @timiagama/ecr lint ./docs    # check structural compliance
+npx @timiagama/ecr stats ./docs   # measure the structure you already have
+```
+
+### Try it on the example corpus
+
+A small [example corpus](examples/docs/) ships with the package, so you can try the tool before
+pointing it at your own documents:
+
+```bash
+npx @timiagama/ecr stats --example
+```
+
+```
+  documents                      9
+    with a References section    9
+  sections                       34
+
+  unique DocIDs                  9
+  unique SectionIDs              34
+
+  References entries             21
+    authority                    5
+    constraint                   2
+    contract                     2
+    dependency                   12
+
+  inline see/per references      17
+    section-precise              10
+
+  total edges                    38
+```
+
+### What the commands do
+
+`init` copies the ECR documentation into the directory you choose. Open its
+README.md for the user guide, specification, navigation protocol and examples.
+The folder structure matches this repository, so the links work locally.
+Keep this documentation outside your own corpus—for example, ECR documentation
+in ./ecr and your documents in ./docs. It also adds that directory to
+`.ecrignore`, so linting your whole project with `ecr lint .` skips it.
+
+`lint` exits `0` when the corpus is clean, `1` when it contains errors, and
+`2` when the command could not run, so it drops straight into CI.
+`--format json` gives you the same diagnostics as data.
+
+READMEs, `CLAUDE.md`, contributing guides and changelogs are excluded
+automatically. Use `--ignore` for anything project-specific:
+
+```bash
+npx @timiagama/ecr lint ./docs --ignore '**/*-CHECKLIST.md' --ignore 'LAST-REVIEW.md'
+```
+
+To keep patterns out of every command, put them in a `.ecrignore` file in your
+project root, the directory you run `ecr` from, one per line, relative to that
+directory. Blank lines and lines starting with `#` are ignored. Folders named
+`node_modules` are always skipped.
+
+A pattern ending in `/**` excludes a whole directory, which is then not read at all.
+
+Symbolic links and junctions inside the corpus are not followed, because `rg`
+and `grep -r` do not follow them either: a document your agent's searches cannot
+reach is left out rather than validated. Each link that could lead to a document
+is listed as not followed; one covered by `--ignore` is listed as excluded
+instead. The directory you name may itself be a link.
+
+There is no `backlinks` command, and there never will be. Navigation is what
+`grep` is for; a navigation command here would make the tool a dependency of
+the thing ECR exists to prove unnecessary.
+
+You can adopt ECR entirely by hand. The tooling is optional.
+
+## Adopting ECR
+
+### Migrating existing documentation
+
+You do not need to convert everything at once. ECR degrades gracefully: a
+partially-converted corpus is more navigable than an unconverted one, and
+documents without DocIDs are simply outside the graph.
+
+A workable order:
+
+1. Assign DocIDs to your architectural documents — the ones that constrain
+   other documents. Leave READMEs and guides alone.
+2. Renumber headings to extend the DocID.
+3. Add `## References` to each document, typing the direction of each edge.
+4. Convert natural-language references ("as described in the storage doc")
+   into `see` / `per` citations with numbers.
+5. Run `ecr lint` and resolve what it reports.
+
+Step 2 changes heading numbers to SectionIDs (`## 8.1#3 - Retry Semantics`),
+and step 4 needs corpus-wide knowledge: resolving
+`as described in the storage doc` to `per 7.1#2` means knowing what documents
+exist and what they cover.
+
+#### Using a coding agent to migrate
+
+Migration always requires knowledge of the corpus, which makes it a good fit
+for a coding agent. Much of the work is mechanical once the DocID set is known,
+but it is too corpus-aware to script blindly.
+
+My advice, from doing this on a real corpus:
+
+- **Start with two fresh sessions from the same model family.** Use one session 
+  to implement and a separate session to review. In my experience, a fresh
+  session reviews what is on the page rather than what the first one intended.
+- **Finish with a review by a model from a different family.** In my
+  experience, a second model family catches assumptions the first one can
+  repeat.
+
+Run `ecr lint` after each pass. The linter catches structural breakage; the
+reviews catch the judgement calls the linter cannot see, such as whether an
+edge is really an `authority` or just a `dependency`.
+
+### Authoring documents with AI
+
+You do not have to maintain ECR structure by hand. Give the [user guide](<spec/v2/2 - ECR - User Guide.md>) and [specification](<spec/v2/1 - ECR - Structural Specification.md>) to a coding agent, desktop harness or general-purpose model alongside the documents it is working on, and it can help create or amend ECR-compliant Markdown.
+
+The linter checks structural compliance. Human review is still needed for the judgement calls ECR cannot make — for example, whether a relationship is really an `authority`, `constraint`, `contract` or `dependency`.
+
+## What ECR is not
+
+- It does not interpret constraint semantics. It exposes structure; you decide
+  what the structure means.
+- It does not evaluate whether a constraint is satisfied.
+- It does not enforce Markdown style.
+- It is not a search engine. Retrieval is exact-match on identifiers, which is
+  the point — there is no ranking to be wrong about.
+  
+
+What distinguishes ECR isn't that its references form a graph. Any linked corpus
+does. ECR makes the structure **stable**, **section-addressable**, **explicit** and 
+**mechanically reversible**, with **typed document-level relationships**.
+
+## Beyond architecture docs
+
+ECR can be useful anywhere an agent needs to navigate a document corpus by
+exact reference rather than similarity. For example, I used the same approach
+in an oil-and-gas engineering PoC where every answer the AI agent provided had
+to cite its evidence. I converted the engineering documents to ECR-compliant
+Markdown so an agent could locate the exact supporting section on demand. The
+same pattern may be useful for standards, policies, legal documents,
+procedures, or other documents where traceable references matter.
+
+## Documentation
+
+This repository contains the ECR 2.0.0 specification and `@timiagama/ecr`, a
+linter that implements it.
+
+- [Specification](<spec/v2/1 - ECR - Structural Specification.md>) — the formal
+  grammar and rules
+- [User guide](<spec/v2/2 - ECR - User Guide.md>) — writing ECR-compliant
+  documents
+- [Design rationale](<spec/v2/3 - Design Rationale - The Section Separator.md>)
+  — why identifiers carry a `#`
+- [Navigation protocol](protocol/navigation-protocol.md) — the instructions
+  you give your coding agent
+- [Example corpus](examples/docs/) — a small example corpus of nine documents
+  across four folders, to try the tool on and to see ECR in use;
+  `npx @timiagama/ecr lint --example` lints it
+
+## Licence
+
+Apache License 2.0. See [LICENSE](LICENSE).
+
+Apache-2.0 includes an explicit patent grant that MIT does not provide.

package/dist/bin.d.ts

@@ -0,0 +1,20 @@
+#!/usr/bin/env node
+/**
+ * ECR Executable Entry Point
+ *
+ * This file runs the CLI unconditionally and does nothing else. It must not
+ * guard itself with a "was I invoked directly?" check that compares
+ * `process.argv[1]` with this module's path: package managers install the
+ * binary as a symlink (`node_modules/.bin/ecr`), Node resolves the module to
+ * its real path, and the two never match, so a guarded entry point silently
+ * exits 0 without running. Keeping the commands in `cli.ts` and the side
+ * effect here removes the need for any such check.
+ *
+ * It is also where supervision is asked for, because this file knows the one
+ * thing a supervised run needs: the script a child process would have to run,
+ * which is this file itself, at its real path rather than through whatever
+ * symlink invoked it. A host embedding the library instead of the command
+ * gets no supervision, and bounds its own work.
+ */
+export {};
+//# sourceMappingURL=bin.d.ts.map

package/dist/bin.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"bin.d.ts","sourceRoot":"","sources":["../src/bin.ts"],"names":[],"mappings":";AACA;;;;;;;;;;;;;;;;GAgBG"}