This build has recorded the same gap in every phase: the ritual material states counts, materials, timings and maṇḍalas, and never states who may do it, what makes it go wrong, or when to stop.
⭐ The last chapter of the Hanumān-Carit supplies all three, in one paragraph.
An eligibility gate — no tantric mantra from a book without permission from a genuine devotee. A contraindication with a stated adverse outcome — celibacy for the duration, with cases "visible even now" of what happens to people who ignore it. And a stopping criterion, arriving from an unexpected direction: the people who actually want liberation send the powers back.1
⚑ No ritual essay in the volume refers to it. ⭐ It is at the end of the narrative section, in a chapter about the later ages, in a book whose ritual material is filed two hundred pages away.
Before the warnings, the chapter sets out the frame they attach to, and it is a fork rather than a ladder.1
Vīra — heroic. For removal of calamity. Rājasika offerings.
Dāsa — servant. For wellbeing and happiness. Sāttvika offerings.
⭐ The mode is selected by the practitioner's situation, not by their attainment. ⚑ Which is unusual: most graded schemes in either corpus sort practitioners by readiness, and this one sorts them by what has gone wrong.
⭐⭐ And it is independently attested inside this same volume — the worship essay arrives at the same two-mode scheme, with the same assignment of registers and offerings, without citing the Carit or being cited by it.2 ⚠ Two contributors converging on a two-mode structure is worth more than either stating it alone, and neither knows about the other. [n=2]
The chapter does something the volume's ritual essays largely do not: it names its sources as books a reader could go and find.1
Anuṣṭhāna-Prakāśa* · Mantra-Mahodadhi · Mantra-Mahārṇava · Mantra-Saṅgraha · *Hanumat-Upāsanā-Kalpadruma
⭐ Five titles, and one of them the build has already worked from directly — the Mantra-Mahodadhi's two vidhis, with their shared sarvatobhadra maṇḍala and their divergent counts.3
⚑ Why this matters more than it looks. ⚠ The build's standing complaint about this corpus is not that it is wrong; it is that claims arrive without chains of custody. ⭐ A closing paragraph that says here are the five books this comes from is doing the one thing the rest of the ritual material does not, and it does it at exactly the point where the stakes are highest — immediately before telling the reader not to attempt any of it unsupervised.
The second warning is the structurally important one, and its precision is what makes it usable.
⭐⭐⭐ *"Do not undertake tantric mantras simply by reading them out of a book; it is useful to proceed only after obtaining permission from a genuine devotee of Rāma or of Hanumān."*1
⚑ Read the gatekeeper specification carefully, because it is not the obvious one.
Not a guru. Not an initiated lineage-holder, not a dīkṣā from a named tradition, not membership of a sampradāya. ⭐ A genuine devotee of Rāma or of Hanumān.
⭐⭐ That is a far lower bar institutionally and a far higher one personally, and the difference is the whole point. A lineage credential can be obtained; the qualification named here cannot be conferred by anybody.
⚠ And it has an obvious weakness the chapter does not address: genuine is unverified and unverifiable. ⚑ The gate has no test attached — which means in practice the seeker selects their own gatekeeper, and a seeker determined to proceed will find a genuine devotee who agrees.
⭐ What survives that objection is the structural claim, and it is worth extracting: the rule forbids acquiring the practice from the text alone. ⚑ Whatever else the permission does, it inserts a second person between the book and the practitioner — and the build has recorded a modern corpus that prescribes 90–180 minutes daily at crisis intensity and names no clinician, no teacher and no second person at all.4
The first warning is blunter and it names consequences, which nothing else in the ritual material does.
"Observe complete celibacy for the duration as far as possible" — and then the author states plainly that there are cases "visible even now" of people who ignored this rule in desire-driven Hanumān worship and "not only failed to obtain what they wanted but were struck with severe physical illness or came under divine displeasure."1
⭐ Three things are specified that this corpus almost never specifies. A rule. A population it applies to — desire-driven worship, not worship generally. And an adverse outcome, in two named forms.
⚠ *The evidence offered is "visible even now", which is not evidence.* ⭐ No case is described, no number given, no mechanism proposed. ⚑ The build records this as a claim of the strongest available type supported at the weakest available level — and the type still matters, because a tradition that can state an adverse outcome at all has a category the rest of its ritual material lacks entirely.
⚠ And the hedge is real: "as far as possible." ⭐ A contraindication with a built-in softener is weaker than one without. The build notes it rather than reading past it.
The third element is the one the build had given up expecting, and it does not look like a stopping rule until you sit with it.
⭐⭐⭐ *"Hanumān is certainly a giver of siddhis… but those who are anxious to cross this wilderness of existence… do not want siddhis at all. They send back the ones that come. They know perfectly well that siddhis are not pleasant but are terrible obstacles to self-realisation… so they stay permanently wary of them, to the point of treating even looking in their direction as an offence."*1
⚑ This is a stopping criterion phrased as a description of a class of people. ⭐ **Not stop when X — but: the practitioners who are actually going somewhere are the ones who return the results.
⭐⭐ Which gives a practitioner something checkable about themselves, and it is the rare kind that can come back empty: when something arrives, do I send it back? ⚠ Nobody else needs to be consulted, and the failure mode it detects — accumulating powers and being pleased — is the exact failure the corpus's own recognition signs cannot see.4
And the chapter states what he himself wants, as a disappointment rather than a rule: *"he does fulfil wishes; he runs the moment a suffering, afflicted, or distressed person calls him… but when he sees them agitated and frantic instead for the fulfilment of the most trivial and perishable worldly cravings, he becomes disappointed and downcast."*1
⚑ A deity who answers petitions and is depressed by most of them. ⭐ That is a stronger discouragement than a prohibition, because it does not forbid anything — it just tells you who you are dealing with.
Three limits, stated because a page arguing that the gap is closed has to survive them.
It is one paragraph against two hundred pages. ⚠ The volume's ritual essays give counts, substances, maṇḍalas, timings and mantras, and this material is not in any of them. ⭐ A safety rail that exists in a book but not at the point of use is not a safety rail. ⚑ The build's finding is not the corpus has adhikāra doctrine — it is that the doctrine exists, once, filed away from everything it governs.
The gate has no test and the contraindication has a hedge. ⭐ Genuine devotee is unverifiable; as far as possible is a softener. Both would be repaired by a single sentence and neither is.
And the stopping criterion is stated about other people. ⚠ The passage describes what serious practitioners do; it does not instruct the reader to do it. ⚑ Converting they send the ones that come back into send back what comes is a one-word change the author does not make — which is the difference between an observation and an instruction, and this corpus makes that slip repeatedly.
Put the eligibility gate where the practice is, not where the theology is. ⭐ This is the single most transferable finding and it is about filing rather than content. ⚑ A contraindication two hundred pages from the technique it governs will be read by people studying the tradition and missed by everyone doing it.
Insert a second person between the text and the practice, and do not over-specify who. ⚠ The rule's low institutional bar is a feature — it does not require a lineage, a fee or an initiation, and so cannot be gatekept by an institution. ⭐ What it requires is that somebody who is not you has looked at what you are about to do. ⚑ Add the test the chapter omits: name in advance what would make you accept a no.
Adopt the return test as the progress marker. ⭐ When something arrives, do I send it back? It is first-person, it can come back empty, and it moves against inflation rather than with it — the three properties the build has spent ten phases looking for in a self-measure. ⚠ And convert it from description to instruction, which the source does not.
State adverse outcomes even when your evidence is weak, and mark the weakness. ⚑ The celibacy warning's evidence is "visible even now" and that is worth almost nothing. ⭐ The warning is still better than the silence in every other ritual essay, because it establishes that failure is possible and describes its shape. Say what could go wrong; label the confidence honestly; do not let a weak citation become a reason to omit the category.
The Carit's closing chapter for the two modes, the five manuals and the three warnings,1 the volume's worship essay for the independent two-mode attestation,2 and three of this build's pages for the ritual material and the modern corpus's safety apparatus.345
Bounded. ⚠ The Carit names five manuals and quotes none of them. The build has worked directly from the Mantra-Mahodadhi material elsewhere in this phase;3 the other four are named here and unread. [PLAUSIBLE — needs corroboration] [RENDERING — verify against Hindi]
⚠ A bounded absence, and it is the page's central claim, so it is stated carefully. Across the 123 teaching essays and the whole non-tattva site read for this build, no ritual or stotra essay refers to this paragraph, and none supplies an equivalent eligibility gate, contraindication or stopping criterion. ⭐ That is a claim about the material read, not about the volume entire, and the five named manuals are outside it.
🔴 A tension the volume creates across contributors. This chapter makes desire-driven worship risky and gates tantric practice behind a second person. ⚑ The volume's own ritual essays supply exactly the desire-driven prayogas — counts, substances and aims, for named worldly results — with no gate and no warning.3 Both are printed. Neither cites the other.
🔴 A second, internal to this paragraph. The chapter says he "runs the moment" a distressed person calls, and that he is "disappointed and downcast" at what most of them ask for. ⭐ A responsiveness with no filter and a stated preference that most petitions not be made. The volume does not reconcile them, and a practitioner cannot act on both.
Open question, and it would be worth a session. ⭐ Do the five named manuals carry their own adhikāra material? ⚑ If they do, the Carit is passing on a gate it inherited and the volume's ritual essays stripped it in transmission — which is a far more serious finding than an author adding a caution of his own. The build cannot tell from here, and the titles are specific enough to check.
Two corpora, both eventually producing safety material, and the difference is where it sits.
The extension corpus's safety framework is genuinely good and it is late. ⭐ A four-part structure — clinical consultation as a precondition, an observable stopping criterion, four warning signs — stated before any practice in the one document that carries it.4 ⚠ And the build has already recorded that it governs one document while a sibling prescribes 90–180 minutes daily at crisis intensity and names no clinician.
This chapter's is early in the tradition and late in the book. ⚑ Same defect, opposite cause. The modern corpus has the rail in the right position within one document and cannot propagate it across documents. The 1975 volume has the rail in the wrong document entirely.
⭐⭐ What that convergence establishes is worth carrying. Neither corpus fails to know about eligibility, contraindication and stopping. ⚑ Both fail to attach what they know to the thing it governs — and in both cases the mechanism is the same: safety material is written by whoever is thinking about safety, and filed with the rest of their writing rather than with the practice.
⚠ This is the third reconstruction of a rail the tradition already had.5 [n=3] ⭐ The pattern is now stable enough to state as a finding about the corpus rather than about any author: this material's guardrails are repeatedly discovered, and repeatedly filed away from their subject.
The Four-Part Safety Framework. That page holds the modern corpus's best safety construction — clinical consultation as a precondition, an observable stopping criterion, four warning signs, stated before any practice. Set beside this paragraph, the two together identify the actual failure mode, which is neither author's competence. ⭐ Both corpora produced adequate rails. Both filed them where the practitioner will not be standing. ⚑ What the pairing yields is a placement rule the build can apply to everything it has read: a safety element's value is a function of its distance from the instruction it governs, and that distance is measurable in pages. ⚠ The modern framework is one document away from the practice that needs it; this paragraph is two hundred pages away. ⭐ Neither is wrong and both are inert — which reframes the whole adhikāra problem from a doctrinal gap into an editorial one, and editorial problems are fixable by someone building a book out of this material.
Two Licences With No Off Switch. That page records the corpus issuing permissions that have no termination condition — a practice authorised with nothing specifying when authorisation lapses. The return test is the off switch, and it works by a mechanism the licence material has no room for. ⭐ A licence terminates on a condition of the practice; this terminates on a disposition toward the results. ⚑ The pairing produces the distinction: when do I stop? has no reliable answer from inside a practice, because the practice supplies the criteria — but what do I do when something arrives? is answerable, because the arrival is an event outside the practitioner's control. ⚠ And it explains why the corpus's licences have no off switch. Every stopping rule it attempted was indexed to attainment, and attainment is assessed by the person attaining. ⭐ Indexing instead to what you do with an unbidden result routes around the whole problem — and the tradition had it, phrased as a remark about other people, at the end of a chapter nobody cites.
Sharpest implication. ⭐ The build's ten-phase finding — that this material states counts and never states limits — is now precisely wrong in a specific and more useful way. ⚑ The limits exist. They are unattached. An eligibility gate, a contraindication with an adverse outcome, and a stopping criterion, all three in one paragraph, in the last chapter of the narrative section, cited by none of the two hundred pages of ritual material they govern. ⭐ For a book built from this corpus, that converts the hardest-looking task — supplying safety doctrine a tradition never developed — into the easiest: moving three sentences to where they belong.
Generative questions.
What does genuine devotee need to become a usable gate? ⭐ The rule's strength is that it cannot be institutionally captured; its weakness is that it cannot be checked. ⚑ The minimum repair is a refusal condition — name, before you ask, what answer you would accept as a no. Without it the gate selects for permissive gatekeepers, which is worse than no gate, because it manufactures a licence.
Can the return test be run before anything arrives? ⚠ Its whole power is that it is triggered by an event you did not choose. ⭐ Which means a practitioner in an early phase has no way to run it, and the corpus's thirty-six-month curriculum gates advancement long before any siddhi is plausible. ⚑ Whether there is an early-phase analogue — some small unbidden result whose return can be practised — is a real question, and it is the kind a practitioner could settle in a season.